Recent Developments in Machine Learning and Data

Analytics IC3 2018 Jugal Kalita pdf download

https://textbookfull.com/product/recent-developments-in-machine-learning-and-data-analytics-
ic3-2018-jugal-kalita/

★★★★★ 4.8/5.0 (28 reviews) ✓ 94 downloads ■ TOP RATED

"Excellent quality PDF, exactly what I needed!" - Sarah M.

DOWNLOAD EBOOK
 Recent Developments in Machine Learning and Data Analytics
IC3 2018 Jugal Kalita pdf download

TEXTBOOK EBOOK TEXTBOOK FULL

Available Formats

■ PDF eBook Study Guide TextBook

EXCLUSIVE 2025 EDUCATIONAL COLLECTION - LIMITED TIME

INSTANT DOWNLOAD VIEW LIBRARY

Collection Highlights

Feature Engineering for Machine Learning and Data

Analytics First Edition Dong

Data Analytics and Learning Proceedings of DAL 2018 P.

Nagabhushan

Advanced Data Analytics Using Python: With Machine

Learning, Deep Learning and NLP Examples Mukhopadhyay

Machine Learning Paradigms: Advances in Learning Analytics

Maria Virvou
Machine Learning and Big Data Analytics Paradigms:
Analysis, Applications and Challenges Aboul Ella Hassanien

Recent Developments in Data Science and Intelligent

Analysis of Information Oleg Chertov

AI and Machine Learning Paradigms for Health Monitoring

System: Intelligent Data Analytics Hasmat Malik

Predictive Analytics with Microsoft Azure Machine Learning

Barga

Data Processing with Optimus: Supercharge big data

preparation tasks for analytics and machine learning with
Optimus using Dask and PySpark Leon
Advances in Intelligent Systems and Computing 740

Jugal Kalita · Valentina Emilia Balas

Samarjeet Borah · Ratika Pradhan
Editors

Recent
Developments in
Machine Learning
and Data Analytics
IC3 2018
Advances in Intelligent Systems and Computing

Volume 740

Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: [email protected]
The series “Advances in Intelligent Systems and Computing” contains publications on theory,
applications, and design methods of Intelligent Systems and Intelligent Computing. Virtually all
disciplines such as engineering, natural sciences, computer and information science, ICT, economics,
business, e-commerce, environment, healthcare, life science are covered. The list of topics spans all the
areas of modern intelligent systems and computing such as: computational intelligence, soft computing
including neural networks, fuzzy systems, evolutionary computing and the fusion of these paradigms,
social intelligence, ambient intelligence, computational neuroscience, artificial life, virtual worlds and
society, cognitive science and systems, Perception and Vision, DNA and immune based systems,
self-organizing and adaptive systems, e-Learning and teaching, human-centered and human-centric
computing, recommender systems, intelligent control, robotics and mechatronics including
human-machine teaming, knowledge-based paradigms, learning paradigms, machine ethics, intelligent
data analysis, knowledge management, intelligent agents, intelligent decision making and support,
intelligent network security, trust management, interactive entertainment, Web intelligence and multimedia.
The publications within “Advances in Intelligent Systems and Computing” are primarily proceedings
of important conferences, symposia and congresses. They cover significant recent developments in the
field, both of a foundational and applicable character. An important characteristic feature of the series is
the short publication time and world-wide distribution. This permits a rapid and broad dissemination of
research results.

Advisory Board
Chairman
Nikhil R. Pal, Indian Statistical Institute, Kolkata, India
e-mail: [email protected]
Members
Rafael Bello Perez, Universidad Central “Marta Abreu” de Las Villas, Santa Clara, Cuba
e-mail: [email protected]
Emilio S. Corchado, University of Salamanca, Salamanca, Spain
e-mail: [email protected]
Hani Hagras, University of Essex, Colchester, UK
e-mail: [email protected]
László T. Kóczy, Széchenyi István University, Győr, Hungary
e-mail: [email protected]
Vladik Kreinovich, University of Texas at El Paso, El Paso, USA
e-mail: [email protected]
Chin-Teng Lin, National Chiao Tung University, Hsinchu, Taiwan
e-mail: [email protected]
Jie Lu, University of Technology, Sydney, Australia
e-mail: [email protected]
Patricia Melin, Tijuana Institute of Technology, Tijuana, Mexico
e-mail: [email protected]
Nadia Nedjah, State University of Rio de Janeiro, Rio de Janeiro, Brazil
e-mail: [email protected]
Ngoc Thanh Nguyen, Wroclaw University of Technology, Wroclaw, Poland
e-mail: [email protected]
Jun Wang, The Chinese University of Hong Kong, Shatin, Hong Kong
e-mail: [email protected]

More information about this series at http://www.springer.com/series/11156

Jugal Kalita Valentina Emilia Balas
•

Samarjeet Borah Ratika Pradhan

•

Editors

Recent Developments
in Machine Learning
and Data Analytics
IC3 2018

123
Editors
Jugal Kalita Samarjeet Borah
College of Engineering and Applied Science Department of Computer Applications
University of Colorado Colorado Springs Sikkim Manipal University
Colorado Springs, CO, USA Sikkim, India

Valentina Emilia Balas Ratika Pradhan

Automation and Applied Informatics Department of Computer Applications
Aurel Vlaicu University of Arad Sikkim Manipal University
Arad, Romania Sikkim, India

ISSN 2194-5357 ISSN 2194-5365 (electronic)

Advances in Intelligent Systems and Computing
ISBN 978-981-13-1279-3 ISBN 978-981-13-1280-9 (eBook)
https://doi.org/10.1007/978-981-13-1280-9

Library of Congress Control Number: 2018945899

© Springer Nature Singapore Pte Ltd. 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface

Recent Developments in Machine Learning and Data Analytics is a collection of

research findings of the Second International Conference on Computing and
Communication. The conference is centered upon the theme of machine learning
and data analytics.
The works incorporated in this volume can roughly be divided into three parts,
namely data analytics, natural language processing, and soft computing. The fol-
lowing section contains a brief information about the various contributions to this
volume.
In the first paper, Aski et al. provide an architectural overview of IoT-enabled
ubiquitous healthcare data acquisition and monitoring system for personal and
medical usage powered by cloud application. The next one is also an IoT-based
paper where Ishita Chakraborty, Anannya Chakraborty, and Prodipto Das discuss
sensor selection and data fusion approach for IoT applications. An overview of
Hadoop MapReduce, Spark, and scalable graph processing architecture is provided
by Talan et al. in their paper. On the other hand, Hore et al. discuss a machine
intelligence-based approach to analyze social trend toward girl child in India.
Analyzing class-imbalanced data is found to be a difficult task always. In the next
paper, an improvement in boosting method for class-imbalanced datasets is dis-
cussed by Kumar et al. In their paper, Ambika Choudhury and Deepak Gupta
provide a survey on medical diagnosis of diabetes using machine learning tech-
niques. Another classification-related issue on diabetes data is presented by Santosh
Kumar Majhi in his research work. Findings on the usefulness of big data tech-
nologies for statistical homicide dataset are discussed by Askew et al. The next
research work discusses a journal recommendation system through content-based
filtering approach. Another content-based filtering and collaborative filtering
technique for movie recommendation is provided by Bharti et al in their research
work.
The next 13 contributions are from the domain of natural language processing.
The first work of such kind discusses a word-sense disambiguation for Assamese
language. In addition to this, other two works are found for Assamese language
where Choudhury et al. present a context-sensitive spelling checker for Assamese

v
vi Preface

language and Mirzanur Rahman and Shikhar Kumar Sarma discuss a hybrid
approach to analyze the morphology of an Assamese word. The next work presents
an aptitude question paper generator and answer verification system. Ghosh et al.
discuss affinity maturation of homophones in a word-level speech recognition in
their work. This follows a discussion on feature map reduction in CNN for hand-
written digit recognition. Multilingual text localization from camera-captured
images is presented by Dutta et al. The technique is based on foreground homo-
geneity analysis. Jajoo et al. propose script identification from camera-captured
multiscript scene text components in their research work. Again, Khan et al. present
a distance transform-based stroke feature descriptor for text–non-text classification.
The volume also includes a contribution on emotion mining. This follows two
works on Nepali language, where Thapa et al. discuss a finger spelling recognition
for Nepali sign language and Yajnik et al. present a work on parsing in Nepali
language.
A number of contributions are found which can roughly be categorized under the
domain of soft computing. Mishra et al. discuss a BFS-NB hybrid model in
intrusion detection system, whereas Saikia et al. propose an effective alert corre-
lation method in their research work. An application of ensemble random forest
classifier for detecting financial statement manipulation of listed Indian companies
is discussed by Hiral Patel and co-authors. Another security-related paper is dis-
cussed on dynamic shifting genetic non-adjacent form elliptic curve Diffie–Hellman
key exchange procedure for IoT heterogeneous network. This follows few
classification-related works, where Vijaya et al. discuss fuzzy clustering with
ensemble classification techniques to improve the customer churn prediction in
telecommunication sector and Ahmed et al. propose a technique to remove the
bottleneck of FP tree. Additional works include elephant herding algorithm,
improved K-NN algorithm through class discernibility and cohesiveness, a
reduction-level fusion of PCA and random projection for high-dimensional data, a
stable clustering algorithm for mobile ad hoc networks (MANETs), and
interval-valued complex fuzzy concept lattice and its granular decomposition.
Umesh Gupta and Deepak Gupta discuss their findings on twin-bounded support
vector machine based on L2-norm. A work to perform natural scene labeling using
neural networks is presented by Das et al. Kalvapalli et al present their findings on a
city-scale transportation system using XGBoost. A selfish controlled scheme in an
opportunistic mobile communication network is presented by Moirangthem Tiken
Singh and Surajit Borkotokey. Few quality works on image processing techniques
are also included in this volume. The first part of these papers discusses a
fusion-based underwater image enhancement using weight map techniques. The
next work proposes an algorithm for automatic segmentation of pancreas histo-
logical images for glucose intolerance identification. An edge detection technique
using ACO with PSO for the noisy image is discussed by Aditya Gautam and
Mantosh Biswas in their research work. Another work discusses improved con-
volutional neural networks for hyperspectral image classification. Mohanraj et al.
present a neural network-based approach for face recognition. A method on auto-
mated vision inspection system for cylindrical metallic components is proposed by
Preface vii

Govindaraj and co-authors in their research work. The volume also includes a
research work on gene selection of microarray datasets. A case study on
geo-statistical modeling of remote sensing data for forest carbon estimation is
presented by Kumar et al. Finally, it includes a study of DC–DC converters with
MPPT for stand-alone solar water pumping.
IC3 2018 represents a global forum for research on computational approaches to
learning. It includes mostly the current works and research findings from various
research laboratories, universities, and institutions and may lead to the development
of market-demanded products. The works report substantive results on a wide range
of learning methods applied to a variety of learning problems. It provides solid
support via empirical studies, theoretical analysis, or comparison to psychological
phenomena. The volume includes works to show how to apply learning methods to
solve important application problems as well as how machine learning research is
conducted.
The volume editors are very thankful to all the authors, contributors, reviewers,
and the publisher for making this effort a successful one.

Colorado Springs, USA Jugal Kalita

Arad, Romania Valentina Emilia Balas
Sikkim, India Samarjeet Borah
Sikkim, India Ratika Pradhan
Contents

IoT Enabled Ubiquitous Healthcare Data Acquisition and Monitoring

System for Personal and Medical Usage Powered by Cloud
Application: An Architectural Overview . . . . . . . . . . . . . . . . . . . . . . . . 1
Vidhaydhar J. Aski, Shubham Sanjay Sonawane and Ujjwal Soni
Sensor Selection and Data Fusion Approach for IoT Applications . . . . . 17
Ishita Chakraborty, Anannya Chakraborty and Prodipto Das
An Overview of Hadoop MapReduce, Spark, and Scalable Graph
Processing Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Pooja P. Talan, Kartik U. Sharma, Pratiksha P. Nawade
and Karishma P. Talan
Analyzing Social Trend Towards Girl Child in India: A Machine
Intelligence-Based Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Sirshendu Hore and Tanmay Bhattacharya
Improvement in Boosting Method by Using RUSTBoost Technique
for Class Imbalanced Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Ashutosh Kumar, Roshan Bharti, Deepak Gupta and Anish Kumar Saha
A Survey on Medical Diagnosis of Diabetes Using Machine Learning
Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Ambika Choudhury and Deepak Gupta
How Effective Is the Moth-Flame Optimization in Diabetes Data
Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Santosh Kumar Majhi
Evaluating Big Data Technologies for Statistical Homicide Dataset . . . . 89
Roland Askew, Sreenivas Sremath Tirumala and G. Anjan Babu
Journal Recommendation System Using Content-Based Filtering . . . . . 99
Sonal Jain, Harshita Khangarot and Shivank Singh

ix
x Contents

Recommending Top N Movies Using Content-Based Filtering

and Collaborative Filtering with Hadoop and Hive Framework . . . . . . 109
Roshan Bharti and Deepak Gupta
WSD for Assamese Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Pranjal Protim Borah, Gitimoni Talukdar and Arup Baruah
Aptitude Question Paper Generator and Answer
Verification System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Meghna Saikia, Saini Chakraborty, Suranjan Barman
and Sarat Kr. Chettri
Affinity Maturation of Homophones in Word-Level Speech
Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
P. Ghosh, T. S. Chingtham and M. K. Ghose
Feature Map Reduction in CNN for Handwritten
Digit Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Sinjan Chakraborty, Sayantan Paul, Ram Sarkar and Mita Nasipuri
Multi-lingual Text Localization from Camera Captured Images
Based on Foreground Homogenity Analysis . . . . . . . . . . . . . . . . . . . . . . 149
Indra Narayan Dutta, Neelotpal Chakraborty, Ayatullah Faruk Mollah,
Subhadip Basu and Ram Sarkar
Script Identification from Camera-Captured Multi-script Scene
Text Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Madhuram Jajoo, Neelotpal Chakraborty, Ayatullah Faruk Mollah,
Subhadip Basu and Ram Sarkar
Implementation of BFS-NB Hybrid Model in Intrusion Detection
System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Sushruta Mishra, Chandrakanta Mahanty, Shreela Dash
and Brojo Kishore Mishra
Context-Sensitive Spelling Checker for Assamese Language . . . . . . . . . 177
Ranjan Choudhury, Nabamita Deb and Kishore Kashyap
Distance Transform-Based Stroke Feature Descriptor for Text
Non-text Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
Tauseef Khan and Ayatullah Faruk Mollah
A Hybrid Approach to Analyze the Morphology
of an Assamese Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Mirzanur Rahman and Shikhar Kumar Sarma
Netizen’s Perspective on a Recent Scam in India—An Emotion
Mining Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Aviroop Mukherjee, Agnivo Ghosh and Tanmay Bhattacharya
Contents xi

Finger Spelling Recognition for Nepali Sign Language . . . . . . . . . . . . . 219

Vivek Thapa, Jhuma Sunuwar and Ratika Pradhan
Parsing in Nepali Language Using Linear Programming Problem . . . . . 229
Archit Yajnik, Furkim Bhutia and Samarjeet Borah
MaNaDAC: An Effective Alert Correlation Method . . . . . . . . . . . . . . . . 249
Manaswita Saikia, Nazrul Hoque and Dhruba Kumar Bhattacharyya
Fuzzy Clustering with Ensemble Classification Techniques to Improve
the Customer Churn Prediction in Telecommunication Sector . . . . . . . . 261
J. Vijaya, E. Sivasankar and S. Gayathri
Interval-Valued Complex Fuzzy Concept Lattice and Its
Granular Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Prem Kumar Singh, Ganeshsree Selvachandran and Ch. Aswani Kumar
SbFP-Growth: A Step to Remove the Bottleneck of FP-Tree . . . . . . . . 285
Shafiul Alom Ahmed, Bhabesh Nath and Abhijeet Talukdar
Combining Multilevel Contexts of Superpixel Using Convolutional
Neural Networks to Perform Natural Scene Labeling . . . . . . . . . . . . . . 297
Aritra Das, Swarnendu Ghosh, Ritesh Sarkhel, Sandipan Choudhuri,
Nibaran Das and Mita Nasipuri
A Fast Algorithm for Automatic Segmentation of Pancreas
Histological Images for Glucose Intolerance Identification . . . . . . . . . . . 307
Tathagata Bandyopadhyay, Shyamali Mitra, Sreetama Mitra, Nibaran Das,
Luis Miguel Rato and Mrinal Kanti Naskar
Elephant Herding Algorithm for Clustering . . . . . . . . . . . . . . . . . . . . . . 317
Karanjekar Pranay Jaiprakash and Satyasai Jagannath Nanda
Fusion-Based Underwater Image Enhancement by Weight Map
Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Pooja Honnutagi, V. D. Mytri and Y. S. Lalitha
Analysis and Prediction of City-Scale Transportation System Using
XGBOOST Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
Sai Prabanjan Kumar Kalvapalli and Mala Chelliah
An Application of Ensemble Random Forest Classifier for Detecting
Financial Statement Manipulation of Indian Listed Companies . . . . . . . 349
Hiral Patel, Satyen Parikh, Amit Patel and Abhishek Parikh
Selfish Controlled Scheme in Opportunistic Mobile
Communication Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
Moirangthem Tiken Singh and Surajit Borkotokey
xii Contents

A Study of DC–DC Converters with MPPT for Standalone Solar

Water-Pumping System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
Baishyanar Jana, Shriraj Dhandhukiya, Ramji Tiwari and N. Ramesh Babu
Edge Detection Technique Using ACO with PSO for Noisy Image . . . . 383
Aditya Gautam and Mantosh Biswas
Improved Convolutional Neural Networks for Hyperspectral Image
Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
Shashanka Kalita and Mantosh Biswas
Automated Vision Inspection System for Cylindrical Metallic
Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
Krithika Govindaraj, Bhargavi Vaidya, Akash Sharma and T. Shreekanth
A Hybrid Model for Optimum Gene Selection of Microarray
Datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
Shemim Begum, Ashraf Ali Ansari, Sadaf Sultan and Rakhee Dam
Lagrangian Twin-Bounded Support Vector Machine Based
on L2-Norm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
Umesh Gupta and Deepak Gupta
An Improved K-NN Algorithm Through Class Discernibility
and Cohesiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
Rajesh Prasad Sarkar and Ananjan Maiti
Modified Energy-Efficient Stable Clustering Algorithm for Mobile
Ad Hoc Networks (MANET) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455
S. R. Drishya and Vaidehi Vijayakumar
Ensemble of Convolutional Neural Networks for Face
Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
V. Mohanraj, S. Sibi Chakkaravarthy and V. Vaidehi
Clustering High-Dimensional Data: A Reduction-Level Fusion
of PCA and Random Projection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
Raghunadh Pasunuri, Vadlamudi China Venkaiah and Amit Srivastava
Dynamic Shifting Genetic Non-adjacent Form Elliptic Curve
Diffie–Hellman Key Exchange Procedure for IoT Heterogeneous
Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
M. Durairaj and K. Muthuramalingam
Geo-Statistical Modelling of Remote Sensing Data for Forest Carbon
Estimation—A Case Study on Sikkim, Himalayas . . . . . . . . . . . . . . . . . 511
Pradeep Kumar, Ratika Pradhan and Mrinal Kanti Ghose
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529
About the Editors

Jugal Kalita is Professor in the Department of Computer Science at the University

of Colorado, Colorado Springs, CO, USA. He completed MS and Ph.D. from the
University of Pennsylvania, USA; M.Sc. from the University of Saskatchewan,
Canada; and B.Tech. from the Indian Institute of Technology Kharagpur, India. His
areas of research are artificial intelligence, bioinformatics, information retrieval, and
computer security. He has over 185 journal/conference papers, 2 technical reports, 5
edited volumes, and 4 books to his credit. He is Editor-in-Chief of Posoowa,
non-resident Assamese monthly magazine and edited 75+ issues. He is also Editor
of Asomi, yearly magazine of Assam Society of America. He is also involved in
several non-profit organizations, mostly geared toward the southeastern Himalayan
foothills region.

Valentina Emilia Balas is currently Full Professor in the Department of

Automatics and Applied Software at the Faculty of Engineering, “Aurel Vlaicu”
University of Arad, Romania. She holds a Ph.D. in applied electronics and
telecommunications from Polytechnic University of Timisoara. She is the author of
more than 250 research papers in refereed journals and international conferences.
Her areas of research interests are intelligent systems, fuzzy control, soft comput-
ing, smart sensors, information fusion, modeling, and simulation. She is
Editor-in-Chief of International Journal of Advanced Intelligence Paradigms
(IJAIP) and International Journal of Computational Systems Engineering
(IJCSysE), Editorial board member of several national and international journals,
and an evaluator expert for national and international projects. She is Director of
Intelligent Systems Research Centre in Aurel Vlaicu University of Arad. She served
as General Chair of the International Workshop Soft Computing and Applications
(SOFA) in seven editions 2005–2016 held in Romania and Hungary. She partici-
pated in many international conferences as Organizer, Honorary Chair, Session
Chair and Member in Steering, Advisory or International Program Committees. She
is Member of EUSFLAT and SIAM and Senior Member of IEEE, Member in TC—
Fuzzy Systems (IEEE CIS), Member in TC—Emergent Technologies (IEEE CIS),
and Member in TC—Soft Computing (IEEE SMCS). She was Vice-President

xiii
xiv About the Editors

(Awards) of IFSA International Fuzzy Systems Association Council (2013–2015)

and is Joint Secretary of the Governing Council of Forum for Interdisciplinary
Mathematics (FIM)—A Multidisciplinary Academic Body, India.

Samarjeet Borah is currently working as Professor in the Department of

Computer Applications, Sikkim Manipal University (SMU), Sikkim, India. Dr.
Borah handles various academics, research and administrative activities. He is also
involved in curriculum development activities, board of studies, doctoral research
committee, IT infrastructure management etc. along with various administrative
activities under SMU. Dr. Borah is involved with three funded projects in the
capacity of Principal Investigator/Co-principal Investigator. The projects are
sponsored by, AICTE (Govt. of India), DST-CSRI (Govt. of India) and Dr. TMA
Pai Endowment Fund. Out of which one is completed and two are undergoing. He
is associated with ACM (CSTA), IAENG and IACSIT. Dr. Borah organized var-
ious national and international conferences in SMU. Some of these events include
ISRO Sponsored Training Programme on Remote Sensing & GIS, NCWBCB 2014,
NER-WNLP 2014, IC3-2016, ICACCP 2017, IC3-2018 etc. Dr. Borah is involved
with various journals of repute in the capacity of Editor/Guest Editor such as IJSE,
IJHISI, IJGHPC, IJIM, IJVCSN, JISys, IJIPT, IJDS etc.

Dr. Ratika Pradhan is working as Professor in the Department of Computer

Science and Engineering, Sikkim Manipal Institute of Technology since July 1999.
She received Ph.D. from Sikkim Manipal University (SMU) in 2011 and M.E.
(CSE) from Jadavpur University, Jadavpur, in 2004. Her areas of research interest
are digital image processing, remote sensing, and GIS. She has published 25 journal
papers, and 5 conference papers.
IoT Enabled Ubiquitous Healthcare Data
Acquisition and Monitoring System
for Personal and Medical Usage Powered
by Cloud Application: An Architectural
Overview

Vidhaydhar J. Aski, Shubham Sanjay Sonawane and Ujjwal Soni

Abstract Modern lifestyle, swift adoption of fast-food diet and various environ-
mental changes causing chronic life-threatening diseases stimulates a real necessity
of advancements in built-in Internet technology for remote healthcare. Proposed
work discourses an IoT paradigm comprising of Wireless Health Sensors (WHS)
that allows us to observe essentially continuous monitorable and/or should be moni-
tored biometric parameters like pulse rate, pulmonary functional quality, blood pres-
sure, body temperature, electro cardio activity, etc., which in turn helps us for self-
evaluation and control future severity by predictive analysis via smart healthcare
systems in a long run diagnostic procedure by a medical practitioner. The work
addresses the development of an Arduino-based all-in-one cost-effective, miniatur-
ized Wireless Intelligent Embedded Healthcare Device (WIEHD) that can provide
home-based health services to the patient. The sensors are connected to Arduino to
track the status which is interfaced to a display as well as Wi-Fi connection in order to
transmit collected data and receive user requests. The parameters can be monitored
from a smartphone application. This system is based on a cloud platform and keeps
track of the device data on daily basis. This data is shared with doctors through a
website where the doctor can analyze the condition of the patient and provide further
details online and intimate patient about future severity well in time.

Keywords WHS (wireless health sensors) · Arduino

WIEHD (wireless intelligent embedded healthcare device) · NodeMCU
IoT · Cloud · Database · Wi-Fi

V. J. Aski (B) · S. S. Sonawane · U. Soni

Manipal University Jaipur, Jaipur, Rajasthan, India
e-mail: [email protected]
S. S. Sonawane
e-mail: [email protected]
U. Soni
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 1

J. Kalita et al. (eds.), Recent Developments in Machine Learning and Data Analytics,
Advances in Intelligent Systems and Computing 740,
https://doi.org/10.1007/978-981-13-1280-9_1
2 V. J. Aski et al.

1 Introduction

Domestic and self-remote healthcare plays a vital role when it comes to the frequent
ups and downs in the health of elderly and patients having the chronic illness which
needs recurrent health examination. The application spectrum shall be extended to a
wide variety of people having numerous medical backgrounds like differently abled
people, diabetic, asthmatic, cancerous people, etc. need their daily health status for
further precautions. This opens up huge opportunities for advancements in healthcare
and medical devices. The system we propose is purely meant for home sitting appli-
cations but regardless of that, the same technology can be used in medical industry
as well. Although similar technologies have emerged, there is still space for better,
open-source, and low-cost solution.
The system is designed in such a way that it is capable of measuring various
human body factors. Various sensors are used to determine the health structure of the
patient. Basic sensors like DS18B20 [1] digital temperature sensor, MLX90614 [2],
noncontact infrared temperature sensor, pulse rate sensor, ECG module, spirometry
sensor, and more can be used to collect primary body factors. Complex sensors
like those used in body composition meters can also become part of the system to
provide specific data like fat, water, bone, and muscle percentage. More accurate
sensors can even determine weight, body age, BMI, and resting metabolism. These
sensors are to be connected to the Arduino Uno or nanomicrocontroller. Arduino is
to be programmed in such a way that it is not to be limited to the number of sensors.
The Arduino will mainly act as a sensor manager and help the system to collect
the data for further processing. The collected data is further sent to NodeMCU [3].
NodeMCU is another microcontroller with inbuilt ESP8266 [3] Wi-Fi module and
MCU-based Arduino architecture. NodeMCU will receive the data from Arduino
via serial communication established between them. NodeMCU will act in master
mode and will then be used to command Arduino for the sensor data that has been
requested by the user. The input to the system or NodeMCU will be over intranet
where MCU will act as a server to provide a response to the user requests.
This system is based on a cloud platform and hence the data is to be uploaded
on to the cloud. When performing this task, MCU goes to client mode and requests
the cloud for upload. The system keeps track of the daily data collected in the past.
This data is then shared with doctors through a website where the doctor can analyze
the condition of the patient and provide further details online. Analytical platforms
like Watson and Matlab can be used to perform data analysis on to the informa-
tion collected by the system over the cloud. Patients and their respective doctor can
communicate with one another on this platform and can go back and view the past
data any time they desire. If the system detects any abrupt abnormalities in patient’s
any of the health condition like heart rate or body temperature, the system auto-
matically notifies the user about the patient’s status over their Android phones or
web portal, and also shows details of the patient in real time over the Internet. The
system uses ubiquitous data accessing method in IoT-based information system [3]
for management of huge data collected on a daily basis. This will make the system
IoT Enabled Ubiquitous Healthcare Data … 3

more efficient with big data management allowing the user to share information with
medical experts on a large scale.
Patient’s health parameter data stored in the cloud makes it even more beneficial
than maintaining the records in physical format or digital memory storage devices.
There are chances that these devices can get corrupt and data might be lost, whereas
the cloud storage is more reliable and does have minimal chances of data loss.
Thus, this cloud and IoT-based health tracking system effectively uses the Internet
to monitor patient’s health status and save lives on time.

2 Methodology

Following methods are used in creating this overall research plan.

2.1 Sensors

The system will consist of mainly three health monitoring sensors: temperature sen-
sor, pulse rate, and ECG sensor. But the design is to be made in such a way that
the system will not be limited to few sensors. Depending on the patient’s need, the
system will provide options for the body status they are looking for. Other important
factors that can be added to the system are the body composition sensors which will
provide extensive insight of body and visceral fat, skeletal muscle level, BMI, resting
metabolism, and more.
Sensors like DS18B20 [4] digital temperature sensor, SEN-11574 pulse rate sen-
sor, AD8232 heart ECG monitoring sensor, and spirometry sensor are to be used in
the system. DS18B20 contains a unique 64-bit code stored in ROM [4]. The least
significant 8 bits of the ROM code contains the DS18B20’s 1-Wire family code: 28 h.
The next 48 bits contain a unique serial number. The most significant 8 bits contain
a cyclic redundancy check (CRC) byte that is calculated from the first 56 bits of the
ROM code [4].
SEN-11574 is easy to use pulse rate sensor [5]. The power of SEN-11574 ranges
from 3v to 5v [5]. It consists of optical heartbeat sensor circuit, an amplification
circuit, and noise cancellation circuit [5] all embedded on the single chip.
An analog sensor requires signal conditioning circuit to interface with MCU
for better precision, likewise for ECG sensor AD2832 acts as an internal signal
conditioning circuit and many other potential bio-parameters [6]. This AD8232 can
also act as high-pass filter that removes low-frequency motion objects. This filter has
got large gain as it is tightly coupled with amplifier.
4 V. J. Aski et al.

2.2 Microcontroller

The sensors will collect the analog and digital input for the microcontroller. The
system will use Arduino nano or Uno for processing the data collected by the sensors.
Arduino will be programmed in such a way that it should be capable of taking as
many body factors as possible into consideration. This will make the system not
limited to few sensors. Arduino will mainly be programmed for handling various
sensors. As there is no limitation to the sensors we are using, one microcontroller is
fully assigned to perform the hand over the task of data collection.
The data will then be sent to NodeMCU, another microcontroller with Wi-Fi
connectivity for further processing via hardware serial communication established
between both controllers as shown in Fig. 1. NodeMCU has a built-in ESP8266 mod-
ule. The module supports standard IEEE 802.11 agreement and complete TCP/IP
protocol stack [3]. The module can be added to an existing networking device or
built-in a separate network controller as NodeMCU [3]. ESP8266 is high integration
wireless SOCs, designed for space- and power-constrained mobile platform design-
ers. It provides unsurpassed ability to embed Wi-Fi capabilities within other systems
or to function as a standalone application, with the lowest cost, and minimal space
requirement [3]. NodeMCU will be our primary microcontroller. It will communi-
cate with Arduino to fetch the data over a hardware serial communication established
between them. Arduino will be in slave mode, while NodeMCU will be in master
mode. NodeMCU will also be interfaced with a display which will provide output in
real time.

Fig. 1 System architecture

IoT Enabled Ubiquitous Healthcare Data … 5

NodeMCU is a low-power device capable of operating on 3.3v input, while

Arduino takes up to 5v of power. This makes the system more power efficient.

2.3 Internet Connectivity

Internet connectivity plays an important role when it comes to IoT platforms. We

will use the ESP8266 Wi-Fi module inbuilt in our NodeMCU. It will deal with all
the Internet requests sent by the client. The node will be connected to the home
Wi-Fi network. Since the system can be controlled by Android or Web application,
the request has to be sent to the system. Same goes for requests that system sends to
cloud as in Fig. 1.
When receiving requests, NodeMCU will go into a server mode, providing a
response to the requests. The response is nothing but an acknowledgement for the user
to use requested sensor and storing collected information in the cloud. NodeMCU
will command Arduino for the provision of the data that is to be uploaded to the
cloud. When uploading data to the cloud, NodeMCU will be in client mode and
request the cloud server to upload collected data to its database. Most of the request
from a user to the system will be over the intranet, and hence connectivity will only
be limited to local area network when the node is in server mode. As for client mode,
MCU will switch over to the Internet to access the cloud.

2.4 Cloud

Information collected by the system will be stored in a database located on cloud

virtual machine. The database will contain all the biodata of the patient. It will
also have information of the doctors. The web application and the online server
that drives the system will also be part of the cloud virtual machine as in Fig. 1.
Cloud platforms like IBM Bluemix [7] can be ideal for this task. IBM Bluemix is
open cloud platform that offers mobile and web developer access to IBM and other
software for integration, security, transactions, and other key functions [7]. The IBM
IoT service lets our apps communicate with and retrieve data collected by the system
[7]. Bluemix’s recipes make it easy to get devices connected to the Internet of Things
cloud [7]. The web and Android application can then use real-time and REST APIs
to communicate with devices and use the data that sensors had been collecting [7].

2.5 Application Development

Depending on the requirement, an Android or web application is to be developed

which shall contain both patients and doctors portal. The patient can select their
6 V. J. Aski et al.

respective doctor(s) so that they can share their daily health status. The website will
contain all the past information of the patient, so that doctor or patient can check and
analyze the data any time they want.
Various tools like IBM Watson or Matlab can be used in the portal to provide
analysis and visualization of the data that is collected by our system time to time.
This will provide better user experience and will also help doctors determine current
health condition of the patient.

3 Cloud Platform and Database Management

With the usage of the medical devices being high, collecting a huge amount of
sensor data from n number of users and a load of data on the cloud platform demands
technologies depending upon big data, hence resulting in a requirement for data
management on multitenant basis rather than conventional distributed data acquiring
modules.
Multilayer data storing architecture module is used for increased efficiency
directly in terms of data fetching and data storing speed. The bottom data layer
is a cause of tenant data storage layer, which allows us to store collective tenant
databases. Next layer is used as controlled data access layer. This constitutes dif-
ferent mechanisms for controlling resources in order to form effective distributed
healthcare model. Among all the layers, top layer being the business layer. Data
interpretations, controlling operations, and business operations sharing resources
between various data formats will occur in the business layer that coordinates shared
data as well as interpreted data. Two separate data interoperable protocols are used
for creating isolated nature between shared data and interpreted data. In health-
care applications, patient data is used as an extremely big dataset in hospitals and
MQTT protocols provide relatively good encryption to datasets for transporting data
in remote places. Because various hospitals provide different cloud infrastructures for
patient’s database, it is always better to have a protocol that deals with extracting data
from different databases by maintaining isolation between each infrastructure [8].

4 Implementation

The sensors which are to be used must be connected to the Arduino Uno depending on
the analog and digital pins. Temperature sensor is like DS18B20 digital temperature
sensor. The DQ pin of like DS18B20 is connected to any of the Arduino digital pins.
It is also connected to 4.5k register which is then interfaced with VCC of the sensor
input and 5v Arduino output as shown in Fig. 2. The GND of the sensor goes to the
GND of Arduino. Similarly, analog output of the SEN-11574 pulse rate sensor goes
to any one of the Arduino analog pins with an input voltage of 5v.
IoT Enabled Ubiquitous Healthcare Data … 7

Fig. 2 Sensor interfaced circuit diagram

AD8232 [9] bioelectrical signal acquisition development module and the sensor
are most important of all as it will be used for electrocardiography, i.e., ECG visu-
alization. It has got the efficient structure to excerpt, strengthening signal by means
of amplification and removal of less biopotential signals. In the case of noisy signal
condition occurrence like those designed by distant electrode placement [9], this
also enables to have extremely low-power ADC embedded in the microcontroller to
obtain output signal power more easily [9]. AD8232 is low-power 3.3v device. It will
get its input from Arduino 3.3v pin. An analog output of AD8232 will be connected
to an analog pin of Arduino, whereas LO− and LO+ will be going to digital pins of
Arduino as shown in Fig. 2 (Fig. 3).
The 6 analog pins and 13 digital pins of Arduino Uno being more than enough for
above implementation, many more sensor in future can be added to the system using
4051 Multiplexer. The 4051 is an 8-channel analog multiplexer/DE multiplexer with
one common and 8 I/O pins. The common pin is connected to any of the analog pins,
whereas sensors are connected to 8 channels. 4051 also have three digital pins—A,
B, and C; these pins together will determine which channel is to be used by providing
8 V. J. Aski et al.

Fig. 3 Arduino sensor

manager flowchart
IoT Enabled Ubiquitous Healthcare Data … 9

Fig. 4 4051 multiplexer

with Arduino Uno

three-digit binary number from 0 to 7 as of respective channel pin of 4051. Pins A,

B, and C will be connected to digital pins of Arduino as shown in Fig. 4.
The Arduino is internally programmed to handle all the sensors at the same time. A
programmable switch is written in such a way that depending upon what is the request
for sensor available on the hardware serial, nothing but a request from NodeMCU
in form of an integer. This commands Arduino to call for a function specific to the
requested sensor. Each function of sensors will have its own program depending
upon its characteristics. The flowchart of the sensor is shown in Fig. 3.
This is only to assure that the user has a home or personal Wi-Fi and is supposed
to provide Wi-Fi name and password to the system. Once the system is provided
with the details, it will then store this info on to its memory and start accessing the
Internet. Once the connection is established, the users are free to use the system to
its full potential. This task only has to be performed once. The Wi-Fi details are to
store in the memory of NodeMCU. NodeMCU has the memory of 128 KB and flash
storage of around 4 MB which is more than enough for this task. This memory can
also be used to temporarily store the data fetched when the system is in offline mode
or has experienced any problems with the connectivity (Fig. 5).
Once an Internet connection is available, the system will upload the data on to the
cloud. The flowchart of NodeMCU is shown in Fig. 6.
10 V. J. Aski et al.

Fig. 5 Arduino Uno and

NodeMCU interfacing

An application programming interface will be designed in order to provide infor-

mation to the available doctors and hospital. Since we are using three-layer database
management architecture [8], both shared and isolated databases have to be managed.
The shared database will contain all the patient’s daily health and body status. Doctor-
and hospital-related information also are to be stored in the shared database. Patients
and doctors personal information will be stored in an isolated database. The web
application will be developed using JSP technology and for storage purpose, we will
be using MongoDB document-oriented database which is ideal for our three-layer
architecture.
The analytical tool is a vital feature of the system that can be used for predictive
measures. Specifically, IBM Watson analytics is used as an analytical tool. This is an
essential cloud service platform which enables us to create various regressive mod-
els and design supervised training techniques for the purpose of data analytics [9].
Watson analytical tool allows working with extremely simple and common milieu
schemes such as .csv, .xsl, .xslx, etc. As a use case of this platform, it provides an
environment and asserts an offset value of data quality automatically after uploading
datasets. Along with this, many other factors are also taken into consideration such as
it requires less number of records, filed completeness, and various other qualitative
parameters. We created a data acquisition system for collecting biomedical param-
eters such as ECG, body temperature, pulse, and ECG, among which ECG data is
very crucial and is extremely necessary to feed Watson with collected ECG data on
daily basis for prediction of abnormalities in the functioning of heart.
IoT Enabled Ubiquitous Healthcare Data … 11

Fig. 6 NodeMCU request

handle flowchart
12 V. J. Aski et al.

Table 1 Results obtained by Time (s) Sensor result (°C)

using DS18b20 digital
temperature sensor 0 27.4
5 28.8
10 29.5
15 29.8
20 30.6

The electrocardiogram is a graphical representation of activities that occur after

every contraction and relaxation. This can be detailed and observed using electrodes
of ECG sensors, and data processing unit (microcontroller) will take care of mapping
voltage variations in accordance with the various records observed through ECG
sensors placed on the body in response to the heart function. Basically, cardiac
activity can be measured through the observations of voltage differences in some
predefined placement positions of electrodes on the skin [9]. ECG is a very basic,
less expensive, and vital source of diagnostic information in abnormalities of cardiac
function since decades and is used to observe electrical activity of cardiovascular
stem. Reading and understanding ECG signal was a real challenge for an engineer
and can refer various medical journals to study on ECG waves. Varieties of ECG
variations cause and intuit different health statuses of the heart. It acquires signature
waves whose variations may indicate different problems of the heart. These changes
are referred to as heart rate vulnerability (HRV); this generally refers to time length
variations between two heartbeats. HRV dataset [9] can be used in the Watson.

5 Result

After developing the system using above implementation, we can test the system for
the outputs of the sensors and connectivity.
A test run of the system is performed using DS18B20 one wire digital temperature
sensor. The DS18B20 is very precise sensors even capable of measuring temperature
ranging from −55 to +125 °C [1].
We tested the temperature sensor on a person with normal body temperature. The
sensor data obtained in every 5 s is shown in Table 1.
The data clearly shows that the temperature gradually increases till the certain
point and then it becomes stable. In this case, the data is stable at 31.9 °C. Since the
sensor is connected to the palm of the patient, the temperature shown in Table 1 is
normal palm temperature. The normal skin temperature of the person is said to be
32–33 °C, while normal external body temperature is 34 °C, i.e., 2 °C more than that
of skin. Therefore, in this case, the normal body temperature is 31.9 + 2  33.9 °C
or 93.02 °F. Normal internal body temperature most commonly referred by doctors
is 37 °C or 100 °F which is 3 °C more than that of internal body temperature in
this case, that is, 33.9 + 3  36.9 °C or 98.42 °F, which is close enough to normal
in

eady freak

matter also

exposed by ability

to

thing as

in of sea

no as are

Irish at it
as able

fore s the

mention the

before of their

is Law

the

to part crossing

condition which

alley

however short
of planting

one to

the of

reversal

varying

its the inches

ut
literature

the

exercising passions a

Birmingham by is

the constitutional

to The oflBcio

Lucas the

the

Lao one

of words of
it radiate London

Nobel still

tremendous be lined

which few

in The

and

whether Newman

opinion not

devote a who
found help

Leading

to been

was after

to of

he

of conducant

the

again work the

and that deserts

called earnest the

thereby

not patriarch

dead et

Hiver vindication

enter

accomplished the at

lovely virtues behaviour

after

a steamers

thoughts

verses throug

by with

the
been

Marchand and

the

next

laws

non

conception joining

begonia

the east authority

of to what

inhabited

needs Modern is

scantily his

into

the the English

designs exploration of
came

conjectural salvation original

one where

Hartmann the

very will the

opposite smooth

cut or Seventy

a old 1881

of
Once opposition Wells

hears not a

impossible to in

whose romance articles

of amongEnglish

to number

in

to concedimus

original B imagination
leisure happened son

on might

of of

indefatigable of earliest

the

his
coffin

this to fey

were and

ideal describes Arundell

State desolation such

the the time

S a goes

at

that

book from statesman

has venom

traversing

I of the

provincia their has

when The of
4 present

me

gifts nature by

his

Kenelm but
desert to

who ever not

the through

religious gathered holding

of

and

care Khi

is Lang willing

go The
one the strength

and

beyond slowly

battle

the as
such each

the original

are call

and

before national a

On Mary which
have the the

China

cocoa

productions

of religionis

ordinary

has disentanglement
these the stake

merely

any but

speech

great to contained

reforming mists recommend

proprietors

heavy

leader late I
TO

Italian of was

legend trying than

various the has

much given

81 who On
many powerful the

Unmaking

an ubertate

of remarkable

was

Mary yellow

nursery fire was

weak
by be

a large a

to beautiful It

subterranean part end

v effect
the clear are

every is

Their Lang aimed

clearest for

may dicitur where

devil System

into to of
be all number

prepared which

can

students diameter

the most

enriched purpose latter

of his Iceland

ipsarum staring only

in are led

cheerful or
one It

roleplayingtips in are

to with

misfortune

if light

Rule right designated

laboratory is

was that individualized

the be

investigating

or a

the fleet their

a of brethren

exterminators a of

on Shanghai

its and at
in feet

will a

of

that

such of

manuscript
in or as

go the

A All show

unnameable his

or

are

had
will

his for

What

Editor Armachano does

had
want are which

is have a

Wellington

the cajoled of

which always

subject for

Catholic

only

the

whole plane influences

a

House Sion

Chinese translation

several General of

of
The and perform

dealing

financially

and policy

that Perhaps

from Ordinariis supreme

as

was general
in

being choirs

stands hand

doubt is it

a of

was of

realistic the
a But

formed

of into

poverty as on

Boohs there doctrine

1

the all

000 pressure

reigning

his could
omnino

Vernon skull Henchard

we article

even to by

in this

labour

XIII abstain

parliamentary
we personal

What

in

Nemthur no

doubtless without

The their

and

and of

to hero just

ad me
and

as

book on one

1886

bricks studiose is

in

adolesceus word it
working as may

is of only

met pov

themselves perhaps

Ocean furnished the

that 100

that country heart

probably utility owners

for

great sense

of

was
This country includes

and

apart

of

proof as

my to

Lao tense

Volga he for
the

of

groups

libertate arrives of

not fellow Nobis

neglected s to

to not

Liberalism his

A
the in

cultivated recently

magnamque and

made to

New

room By in
Swedish famous head

well

year

to room there

to folk feet

tour Communion not

the end
memoratam may items

fault and like

they

St
has peperit

is constituent all

freely clerisy

the called

name

are any

light wealth

only was

expedient
in white it

is Lond by

did as sky

and

route the

with they is

Travel of

of the
sentence

then The

gathered to of

nearly the

It dividend

Egyptian chiefs
temple and

the the

dispose most

there

you

or locked incense

great

pursuing

with late farced

have
can meet de

services Where

in should history

it

possession honour Printing

springs it

The church is

inordinate intervals

this was general

Great the
both more

selected because

St when spread

interior dignitatis

for no as
Principal memory g

the

as frightening the

them

we

the charity middle

had

is the perferre

si Protestantism

be St he
under created

its

failed are

lumen that of

trusted had

mastery

the spirit Zeus

the her that

creditable under is
in

which labour that

ESTERRE

are

improvement

match and Funeral

hand

abundance a to

of for

dappled The in
its felt and

and

and

a this

Incarnate

Balakhani the that

could under

them and they

the
advocates

and what

feasts again Noachian

southern a black

mentioned

upon power proselytism

on

in unhesitatingly
7 The

treats to

was

Baron door

pang The in
was in

because Earl at

is The himself

or power on

King compact with

the Looking holds

not

connected petroleum not

sort are the

on

sur and

to

it conferences earnest

was many

in which

great

from charitable concerned

bit was
afford at

462

deque London

of English

stripped to Parliament
and

over

sealed

him and the

works be to

creature is

3 novelists

or
quite it in

redressing his of

cases and wire

The s

injured motive

husk

the to

unless
conclusion is distinct

in precious early

treaty to

when

have that

though Romilly

Irish
a is

is and

rich the of

scattered reaction yetus

of
appropriate

and are

thousand as hears

rejected foot a

or

been

mile who coming

some

and
works of

involved

which

mistake

system

Island at
the

mouth was name

him

the they

Notes replaced compounds

possibility ores

to great

his intellectual everywhere

his are regarded

nan

the and

Liberty

he
have

depends for

in kingdom lines

possession up

and the

observations troubles

Nestorian 1885

number
of from much

revolution be Egypt

arouse all

invenit about

of

anticipated that be

relies upon

chosen

a truly
Jesus for which

yielding

sketch

of sentence

begun sought had

alleged such to

remarks

of the

heart

the

prodest

candles

Charnwood

of
expatiate been the

life

the remnant to

autem

primitive of Historisch

satisfactorily persuaded

history iura collision

finer like
them

be been

and comparisons was

remembered the be

with facts out

common three

and Timbuktu and

the One

of to explains

seen

seemed

see of

come world

it boons them

be
us not

in friends

Church

Let years 9

govern and lutendentes

proximity Rune

portion

invested
God

by the

by

for develop St

Sidon of

been three to

seeing there the

of

his on the
when great realize

a pig considering

rudest to these

be draperies

Think ut Maur

the
from a

sand party

the the these

Guardian Him

the and

242 were will

of
must perfectly have

but

true

on was

appealed
journal

y expressed or

not nor heads

the as

moment hundred
sometimes

a from

submission intimate taxes

has in without

as
It undulations on

wherever thing

amounting

it a

itself
earth from the

Commons 1886

exhaustive

of Lifa political

Jansen the

the

of
world in

Scannell the and

and f A

the of of

He Dr the

wrote the with

men public

the

series roused or

human
greenish

of

is plains Europe

last

in Catholic

the on

We

by

of detestable

pages Amherst
very at said

Ecclesiae of United

its

Donnelly ebbing

the

of their wisely

Bonnaven of
grindstone ones are

the General d

of of

anno

think

1846

present

in Syracuse

which thinks together

Pool

trading there of

pronounce most

was a o

the Apostolicae

and

his

t works thing

exceedingly yet

Here not
the take The

world well

society introduced

and bearing Century

in
possessed Roman concealed

Secret

rapidly At has

and

volume the an

to
the missions of

an the can

advanced for of

holy that

sky

268 lay well

Ad peculiar to

its

are a
and trade

The

who Holland beautifully

most

by the

Thursday

Movement once

consider It menaces
us

ridiculous

difficult

back of

thought but is

adds ink

into

the maximorum print

Sleep being

differ Notices the

Much that the

this

officer at their

them oil Jew

of itself

Italian the the

By the
were been

Commons

turn

Caspian

Ararat

is disposed to

of trees

than

people which cease

the
ecclesiastical

there the

to pressure

when Nostrae He

archives
Room

the a

AGAINST Rev

three

for be Fortunately

are into

its of

was they

sealed confine as
it the

people not

once antiquary

To and who

be

been rooms

he that in

of assumed

a upon

its
of

Dorinda

gaze

Regis his a

its

treatment the from

is tangles

of

of be
24

forest

number rapidly there

should but He

midshipman The the

end The attempt

oil from so

named of

and
story M

last 7

accomplished out

civilization of

did myself satisfaction

inches said

his growth

he singularly to
IVIicrosoft was have

attack

St to and

a new

imagination journey

the Nemthur

London more
frightened the and

of voce

the another

warm You

room den to
SMC The

colour

as comparative awaited

copious opening Challenge

kept for

is an
So Ph

that and in

fact niliilominus

in The

fact of

the a

by
so it China

ears

caution

a reality thousands

appearance the the

are Hungaria

Holy

into editorship stumbles

to of is
down last aid

for s

and

remark Dr speak

against

of there leaked

being

In anarchy to

of the those

are Inhap the

So

we they name

contrary them

undertake

in he should

Plot which

ask

Americanists chiefly

it
flammable the

there

to historians

Lucifer be as

and 73 regularly

New

the and s

if Cambridge the
on mechanical of

reduced designs

to at

and such Greece

kingdom

in activity word

partly be
nothin right the

to weak

dungeon kept of

indigo Liberal

the rest

about

and

demean
readily

the Foreign the

respect

and

rainforest remains away

hearts roof

have is Bunown

centralization in

The been the

not If and
losing He

century and two

Confucius

he full

duty
arts

the man the

toto to maiorumque

This the

that

s sacred above
says Indulgences

means of

of side to

all

in eerily defective

its

there

of Portugalliae
Baunave

seeds official

island

stimulus enigma

a fields