Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

Development of Faculty Qualification Analysis

System using Naive Bayes Algorithm
Eliza B. Ayo1; Lani E. Sakay2 Anna Liza O. Villanueva3; Rosalina R. Pangilinan 4
Centro Escolar University, Manila, Philippinnes Perpetual Help College-Manila, Manila, Philippines

Celia L.Verano5
Manila High School, Manila, Philippines

Abstract:- Qualification prediction is a crucial process in The current practice involves a manual review of printed
determining whether an applicant is qualified for a curriculum vitae, which can be time-consuming as both parties
particular position. However, traditional methods of need to thoroughly read through all the criteria. This manual
evaluation often rely on the experience and intuition of the approach occasionally leads to conflicts arising from differing
evaluator, which may not always be accurate. This study interpretations of standards or variations in awareness of
proposed the use of a supervised machine learning specific requirements. For some institutions, the search
approach, specifically the Naïve Bayes algorithm, to committee's faculty votes are more influential in tenure-track
predict faculty qualification based on a labeled dataset. faculty selection decisions than the chair or dean's votes, with
The developed Faculty Qualification Analysis System for academic accomplishments, interview performance, and
Perpetual Help College of Manila would allow users to presentation skills being key determinants [22]. Others opted
input appropriate test data and generate results of for AI adoption in their recruitment process because it leads to
qualified or not qualified. The system’s effectiveness and efficiency and qualitative gains for both clients and
acceptance had 4.3 and 4.4 ratings with verbal candidates, offering strategic insights into automation and AI
interpretation of very high and strongly acceptable. The implementation [23]. These technologies have the potential to
results of this study demonstrated the potential of machine save time and resources while improving the accuracy and
learning algorithms to improve the accuracy and efficiency of recruitment. In this study, the naive Bayes
efficiency of qualification prediction processes in algorithm improves the traditional method by selecting a
educational institutions. subset of attributes, enhancing classification accuracy while
reducing computational overhead [1].
Keywords:- Qualification Prediction, Machine Learning,
Naïve Bayes Algorithm. Similar to this is the use of Artificial intelligence in
Indian software companies which positively impacts the
I. INTRODUCTION recruitment process resulting in effective talent acquisition
and sustainable development [24]. The study is expected to
As educational institutions encounter heightened help organizations formulate recruitment strategies and policy
competition in attracting top-tier faculty members, Human interferences to align to develop its effective recruitment
Resources (HR) departments face the imperative of process to recruit qualified talent into the team to encounter
implementing efficient screening procedures to identify the competitive business and to develop a sustainable
most qualified candidates. Perpetual Help College of Manila is environment. Moreover, several educational institutions also
no exception. Its Human resource department must scrutinize have adopted predictive algorithm-based screening systems
applicants' credentials to ascertain compliance with the like, a college faculty recruitment system [25], an android
college's requirements before proceeding to demo teaching application that helps in sorting out candidates and providing
sessions. Nevertheless, the prevailing method of manually the best results in the recruitment cycle, and a comprehensive
reviewing printed curriculum vitae can prove time-consuming campus recruitment and placement system for optimizing the
and could introduce conflicts or errors during the assessment hiring process [26].
process. To overcome this challenge, this study is anchored on
the development of a Faculty Qualification Screening system With the mentioned successful implementation, a shift in
that uses machine learning and predictive analytics. This paradigm to ditch traditional methods of faculty recruitment
innovation can free HR personnel to do responsibilities and process is necessary. Adopting a machine learning and
devote their time to other important tasks. predictive algorithm-based screening system can streamline
processes, and save time, to ensure HR hires qualified
In the case of Perpetual Help College- Manila, before a candidates is necessary. With the help of machine learning
faculty member conducts a demonstration of teaching skills, using the naïve Bayes algorithm, automated faculty
the Human Resource Department needs to assess whether the recruitment is possible. The developed systems allow greater
applicant's credentials align with the university's faculty diversity to create a richer learning environment and
requirements. Collaboration between the Human Resource ensure continuity to be able to recruit top talent which
Department and the Dean is integral in the evaluation process improves the overall quality of education provided at the
to determine the eligibility of candidates for demo teaching. institution.

IJISRT24FEB1647 www.ijisrt.com 1719

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

II. STATEMENT OF THE PROBLEM This study employs the predictive modeling theory as
illustrated in Figure 1, with a specific focus on supervised
 This Study was Pursued to Answer the Following machine learning and the Naïve Bayes algorithm. The process
Problems: begins with data collection to gather pertinent information for
the subsequent modeling stages. The feature selection follows,
 What are the challenges encountered in the existing involving the identification of crucial variables that impact the
screening process for faculty applicants to be qualified for model's outcome. Subsequently, an appropriate model is
teaching demonstrations? selected to ensure accurate predictions based on the chosen
 How is the Naive Bayes Algorithm used in the data sets to features. The model is then trained using a designated training
predict qualified faculty applicants? set, and its performance is evaluated through carefully
 How does the developed prototype address the challenges selected test data and diverse metrics, parameters, feature
found in the existing system? selection, or preprocessing techniques to enhance its efficacy
 How do the I.T. experts assess the level of effectiveness of and optimization [27]. Once optimized, the model is ready for
the developed prototype in terms of ISO 25010 criteria deployment. This comprehensive workflow diagram provides
a systematic framework for constructing and implementing a
 Functional – Correctness; supervised machine learning model for predictive purposes.
 Efficiency/Performance -Software Capacity;
 Usability – user error protection;  Assumptions of the Study
 Reliability – fault tolerance and recoverability; The current process for determining qualified faculty
 Compatibility; applicants for teaching demonstrations has deficiencies and
 Security; inefficiencies that can be addressed through the developed
 Maintainability; and system.
 Portability?
 Scope and Delimitation
 How do the users assess the acceptance of the developed The study focuses on the development and evaluation of
prototype in terms of ISO 25010 criteria: a faculty qualification analysis system utilizing the Naïve
Bayes algorithm at the College of Computer Studies-Perpetual
 Functional suitability – Completeness and Help College of Manila. The key areas of emphasis include the
Appropriateness; screening process employed by the HR department for
 Time-behavior and resource utilization; selecting faculty for teaching demonstrations, creating of
 Appropriateness, learnability, operability, and user training data set, designing the system's database and
interface aesthetics; and lastly architecture, evaluating the system's effectiveness, and
 Maturity and availability? assessing user acceptance. Several limitations are inherent in
this study. Firstly, the use of historical data from the Perpetual
III. THEORETICAL FRAMEWORK Help College of Manila and a Kaggle dataset may affect the
results to other institutions or disciplines when implemented.
The study does not evaluate hardware equipment for system
implementation, and the absence of an exploration of ethical
implications associated with machine learning in faculty
selection, are further limitations. Despite these constraints, the
study contributes valuable insights into the development and
implementation of a faculty qualification screening system
using the Naïve Bayes algorithm at Perpetual Help College of
Manila, with a need for cautious interpretation due to the
identified limitations.

 Significance of the Study

This study significantly impacts the faculty selection
process within academic institutions, aimed to enhance
efficiency, accuracy, and the selection of qualified candidates,
ultimately fostering excellence in academia through the
development of faculty qualification based on the Naïve
Bayes Algorithm. This streamlined approach holds the
potential to optimize resource allocation, make informed
decisions about faculty appointments, and improve the overall
effectiveness of the hiring process. Future researchers can
draw upon the study's findings, methodologies, and
challenges, using them as a stepping stone to contribute
further to the field.

Fig 1 Supervised Machine Learning Process Model

IJISRT24FEB1647 www.ijisrt.com 1720

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

IV. RESEARCH METHODOLOGY The project utilized specific software tools for its
development, operating on the Windows 10 platform. The
The research design and methodology employed in the primary programming language employed was Python, known
study used the descriptive developmental method. The for its simplicity, readability, and support for modularity and
descriptive method's primary focus is to describe, compare, code reuse. The Spyder IDE, included with Anaconda, served
analyze, and interpret existing data, aligning with the study's as the integrated development environment for editing,
objective of identifying criteria for predicting the faculty testing, and debugging the Python code. Additionally, CSV
qualification system and evaluating the software results. The files were utilized for data storage.
development Faculty Qualification Analysis System using the
Naïve Bayes Algorithm aimed to develop and implement an The user acceptance and IT expert evaluation were
efficient system for analyzing the credentials of faculty integral components of the project design, following the
applicants for demo teaching. The design and methodology ISO25010 standard. The evaluation considered factors such as
involved several key phases, data collection and preparation, the refined Naïve Bayes Algorithm, the system's functionality,
training data selection, algorithm development, system and its efficiency in predicting faculty qualification. The
development tools, evaluation and testing, and project project's deliverables encompassed various elements,
deliverables including data collected from printed curriculum vitae and
historical Kaggle datasets, a refined Naïve Bayes Algorithm, a
The project commenced with data collection and fully functional Faculty Qualification Analysis System, and
preparation. The researcher gathered both primary and comprehensive reports on user acceptance and system
secondary data, showcasing the current business process of effectiveness as evaluated by IT experts. These deliverables
faculty screening of Perpetual Help College of Manila. A collectively reflected the successful implementation of the
proposed system was designed based on the representation of designed project.
the data flow of the current system. This design aimed at
creating a more efficient and faster system, objectively  Sources of Data
reducing processes and generating results. To enable the The data for this study was sourced from multiple
system to predict qualified faculty for teaching demonstration, channels, employing a mix of primary and secondary data.
the researcher identified a suitable training data set. The The primary sources included interviews and focus group
Kaggle data set, considered a foundational resource in data discussions, providing firsthand insights and opinions from
science, was chosen to provide historical data for the system's relevant stakeholders and among four (4) IT experts, four (4)
machine learning training phase. The study extensively College Deans, and two (2) HRD staff members. This strategic
explored and employed the Naïve Bayes algorithm for sampling approach allowed for targeted insights from
predicting the probability of faculty qualification. The individuals directly involved in the faculty qualification
algorithm development phase involved designing the system, screening process, ensuring a focused and relevant data
including the database, user interface, and logic design. The collection process. The gathered data from these interactions
features of the Faculty Qualification Analysis System were underwent thorough analysis, leading to the formulation of the
carefully deliberated based on the expected project output. study's objectives. The historical data were extracted from the
The development process adhered to a structured and Kaggle dataset, a repository of information for machine
systematic approach based on the Waterfall Model, a learning training while the data from the survey conducted
well-established model in the System Development Life Cycle was used to gauge the system's effectiveness and acceptance.
(SDLC). Figure 2 provides a visual representation of the Collectively, these diverse sources of data contributed to a
various steps undertaken in the development process. holistic and informed approach to conducting the study.

 Statistical Treatment of Data

To measure the user acceptance and system effectiveness
of the Faculty Qualification Analysis System, a weighted
mean approach was employed. The eight software
characteristics and their sub-characteristics were given equal
weights, resulting in the weighted mean being equivalent to
the arithmetic mean or sample mean. This allowed for a simple
assessment of both factors as seen in table 1.

Table 1 Measurement Ratings - Level of Effectiveness

Assigned Numerical Categorical Verbal
Point Ranges Response Interpretation
5 4.51 – 5.00 Strongly Agree Very High
4 3.51 – 4.50 Agree High
3 2.51 – 3.50 Neutral Neutral
2 1.51 – 2.50 Disagree Low
1 1.00 – 1.50 Strongly Very Low
Disagree
Fig 2 Waterfall Development Method

IJISRT24FEB1647 www.ijisrt.com 1721

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

The acceptance level of the proposed system was both extraction and cleaning processes. Extraction involves
evaluated using ISO/IEC 25010:2011, which utilizes a the identification and selection of pertinent data for analysis
five-point scale to measure its level of acceptance. This while cleaning entails eliminating redundant or irrelevant
evaluation helps determine how well the system meets the information, rectifying errors, and converting the data into a
desired criteria, as presented in Table 2. required format. Subsequently, a comprehensive dataset is
created, incorporating both training and test data, it served as a
Table 2 Measurement Ratings - Level of Acceptance model for accurate prediction and analysis. The training
Assigned Numerical Categorical Verbal dataset plays a crucial role in instructing the machine learning
Point Ranges Response Interpretation model because it will rely on identified patterns and
5 4.51 – 5.00 Strongly Strongly relationships within the data to recognize trends. Following
Agree Acceptable the model training, an evaluation phase follows, where the
4 3.51 – 4.50 Agree Acceptable accuracy and effectiveness of the model are assessed by
testing it on a separate portion of the dataset, specifically the
3 2.51 – 3.50 Neutral Neutral
curriculum vitae dataset. This validation step ensures that the
2 1.51 – 2.50 Disagree Unacceptable system's predictions are aligned with the actual values stored
1 1.00 – 1.50 Strongly Strongly in the database. Finally, the Naïve Bayes algorithm is applied
Disagree Unacceptable to the prepared and validated dataset, enabling computation
and analysis to determine the qualification status of faculty
 Instrumentation and Validation members. This comprehensive system architecture guarantees
The researcher utilized the international standard a methodical and robust process, encompassing data
ISO/IEC 2510:2011 to assess the acceptability of the software. acquisition, preprocessing, dataset creation, and algorithmic
This standard outlines eight key software characteristics, utilization for accurate faculty qualification prediction.
namely efficiency, compatibility, suitability, reliability,
maintainability, security, usability, and portability. These  Data Gathering Procedures
characteristics served as the basis for evaluating the software's For this study, a comprehensive data-gathering
quality, ensuring that it met recognized industry standards. procedure was employed to ensure the acquisition of relevant
and valuable information. The process included the collection
 System Architecture of data from diverse online sources, such as studies, research
papers, documents, and articles, with a specific focus on the
proposed Faculty Qualification Analysis System and the
application of the Naïve Bayes algorithm, as outlined by Putra
et al. in 2020. To assess the effectiveness and user acceptance
of the system, a survey method was chosen as the primary data
collection tool. The researcher meticulously designed survey
questions aligning with the standards set by ISO/IEC
25010:2011, reflecting various software quality
characteristics. Subsequently, the researcher personally
administered the questionnaires to the identified participants.
The survey responses were systematically collected and
documented for further analysis. The collected data underwent
rigorous statistical analysis to draw meaningful conclusions
and insights. This analytical process allowed the researcher to
evaluate the system's effectiveness and user acceptance,
providing a robust foundation for formulating conclusions and
offering recommendations for the study. The combination of
online source analysis and survey data collection ensured a
well-rounded approach to gathering information for the
research study.

 Testing Procedures
The Black Box Testing for the Faculty Qualification
Analysis System was employed. This aimed at evaluating the
software's functionality from an end-user perspective without
delving into the internal code of the system. The testing
Fig 3 System Architecture process adhered to the system's requirements and expected
functionality, focusing on diverse aspects such as the user
The system architecture employed in this study adheres interface, APIs, database, security, client/server applications,
to a systematic process as seen in Figure 3. The initial step and overall system functionality. Further testing measures
involves acquiring historical data, which forms the were implemented, including Install/Uninstall testing. This
foundational basis for the analysis. Following this, the data comprehensive approach ensured a thorough evaluation of all
undergoes a meticulous preprocessing phase that includes system components to guarantee their excellent performance.

IJISRT24FEB1647 www.ijisrt.com 1722

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

 Qualified Output – Evidence 1

Fig 4 Data Flow Diagram

B. Training Data Set to be Used in Predicting Qualified-

Faculty Applicants
 Not Qualified – Evidence 2 The initial stage involves preprocessing the data,
wherein relevant information is extracted and cleaned for
subsequent analysis. Extraction entails identifying and
selecting the necessary data, while cleaning involves
eliminating irrelevant or duplicate data, rectifying errors, and
formatting the data appropriately for analysis.

Following this, a dataset comprising both training and

test data is generated for future use. This dataset serves as a
model capable of accurately predicting and analyzing data.
The training dataset is employed to teach the machine, relying
on patterns and relationships within the data to identify trends.
Once trained, the model's accuracy and effectiveness are
assessed by testing it on a separate portion of the dataset,
specifically the curriculum vitae. This dataset is utilized to
validate the system's accuracy in producing predictions by
comparing the results to actual values in the database.

Upon dataset creation, the Naïve Bayes algorithm can be

applied to compute and analyze the data, determining the
V. ANALYSIS AND FINDINGS qualification of faculty members. The algorithm evaluates and
predicts faculty qualification based on the patterns identified
A. The Current Screening Process of the Perpetual Help during training.
College of Manila for Faculty Applicants to be Qualified
for Teaching Demonstration  The Naive Bayes Algorithm
The Perpetual Help College of Manila (PHCM) currently The Naive Bayes algorithm is a widely employed
conducts teaching demonstrations after the College Dean and method in machine learning, particularly for classification
the HR Department review the curriculum vitae. These two tasks. Grounded in Bayes' theorem, it computes the
stakeholders collaborate to assess if the applicant's credentials probability of a new data point belonging to a specific class
align with the college's requirements. Then the applicant will based on its features. This algorithm calculates the conditional
proceed to the next steps. However, the manual review of probability of each feature for the given class and integrates
printed curriculum vitae has led to various challenges, them to ascertain the overall probability, facilitating the
including: (1) a delay of 3 days of forwarding CVs to the prediction of a data point's class based on its features.
college dean, attributed to the HR secretary's busy schedule.
(2) a 2-day duration for the dean to review CVs due to priority To delve into the mechanics of the Naive Bayes
obligations; (3) subjective recommendations on qualification, algorithm, the initial step involved transforming the provided
influenced by biases and subjective judgments; (4) instances dataset into a frequency table. This table is instrumental in
of CVs being lost due to mishandling; and (5) potential determining the probabilities associated with the provided
conflicts arising from considerations of standards, awareness features. Subsequently, applying Bayes' theorem, we compute
of requirements, or inconsistencies in evaluation criteria. the posterior probability, enabling the estimation of the
Figure – shows the data flow diagram of this process. probability of a specific class given the observed features.

IJISRT24FEB1647 www.ijisrt.com 1723

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

For a more concrete illustration, let's consider an outcome. This helps us determine the most likely class for the
example from javatpoint.com. Suppose we have a training given data based on the calculated probabilities. The training
dataset comprising weather conditions and a corresponding data set is to be used in predicting qualified faculty applicants.
target variable "Play," aiming to predict whether a player To come up with a training data set the researcher went
should play or not. The dataset is structured is presented in the through the process of cleaning and interpretation of data.
table below. Table 7 is the result.

Table 3 Example of Dataset Table 7 Data for Filtering

 To Solve this Problem, Convert the Data set into

Frequency as Shown after this Statement.

Table 4 Frequency of Data

 Then Create a Likelihood Table as Shown in Table 5.

Table 5 Likelihood Table

C. Faculty Qualification Analysis System Designed for

Perpetual Help College of Manila
Table 6 Prediction Outcome The Faculty Qualification Analysis System was
P(Y|S) P(N|S) developed to determine the eligibility of an applicant for
P(Y|S)=(P(S|Y)*P(Y)) / P(S) P(N|S)=(P(N|Y)*P(N)) / P(S) teaching demonstration. Employing a filtering technique
P (S|Y) = 3/10 = 0.30, P (S|N) = 2/4 = 0.5, commonly found in recommender systems, the system utilizes
P(S) = 5/14 = 0.36, P(S) = 5/14 = 0.36, the Naïve Bayes Classifier for classifying large datasets,
P(Y)= 10/14 = 0.71 P(N)= 4/14 = 0.29 ensuring precise predictions for class variables with no parents
P(Y|S)=(0.30 * 0.71) / 0.36 = P(N|S)=(0.50 * 0.29) / 0.36 = due to its exact learning augmentation. The chosen algorithm,
0.60 0.40 based on Bayes' theorem, is known for its simplicity and
Since P(Y|S) is higher than P(N|S), therefore Planer can play effectiveness. It demonstrates proficiency in handling
the game on a Sunny day. high-dimensional data, a challenge for many other machine
learning algorithms. Notably, the role of machine learning in
Lastly, we apply the Naive Bayes equation to compute providing recommendations, as seen in job recommendations,
the posterior probability for each class. The class with the aligns with the system's purpose of predicting qualified
highest posterior probability is considered the predicted faculty.

IJISRT24FEB1647 www.ijisrt.com 1724

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

Table 8 Sources of Data Upon evaluating the historical data, the researcher must
undertake a crucial step: data conversion, depicted in Table 9.
This process is essential to enable the Naive Bayes algorithm
to effectively analyze and process the data, ensuring accurate
predictions by appropriately transforming each attribute for
the algorithm. After the data conversion process is
successfully executed, the training data is now ready for the
next stage, where it can be employed to predict the desired
outcomes. With the completion of the data conversion
process, the training data is primed for the subsequent step,
involving the prediction of probabilities using the Naive
Bayes algorithm.

Table 10 Loook up Table

As indicated in the project development phase, following

the compilation of pertinent faculty attributes essential for the
qualification screening, the next step involves obtaining
historical data. This historical data plays a crucial role as it
Table 10 illustrates a lookup table that contains tabulated
serves as the basis for creating a training dataset utilized by the
data showing the frequency counts of an attribute. By
Naïve Bayes algorithm. Table 8 showcases the data sourced
employing Bayes' theorem, the conditional probability
by the researcher from www.kaggle.com, serving as the
(P(c|x)) can be calculated using the prior probability (P(c)) and
project's historical data and forming the foundation for
the likelihood (P(x|c)). The following equation represents this
constructing the training dataset. Furthermore, the researcher
conditional probability:
meticulously processed the selection of pertinent attributes
from applicants' curriculum vitae, which were subsequently
incorporated into the prediction model for faculty
qualifications

Table 9 Attributes

Fig 4 Numerical Equation

IJISRT24FEB1647 www.ijisrt.com 1725

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

In Table 12, shows the result of the computed values of  To Provide Another Test on the not Qualified Result,
the probabilities for sample Evidence 1 P(x). The P(x) are the Probability of Evidence 2 is Presented as Shown in Table
attributes that serves as the predictors. Using the formula of 13.
Naïve Bayes theorem and the use of Look-up table in Figure
11 with the given attribute condition, the probability of Table 13 Probability of Evidence 2
qualified and not qualified is computed.

Table 11 Probability of Evidence 1

Based on Table 12, P(nq|A) the probability of not

qualified faculty with the given predictors is higher than the
Based on Table 11, P(q|A) the probability of qualified
faculty with the given predictors is higher than the P(nq|A) P(q|A) probability of qualified faculty. Therefore, the
applicant is not qualified as shown in Table 11.
probability of not qualified faculty. With the given faculty
applicant attributes of high seminar, very good certificates,
Table 14 Probability of Evidence 2 Result
high training, more than three years of experience, excellent
educational background and specialization in Computer
Programming, the applicant is qualified with P(q|A) of 0.0376
and an equivalent average of 65.53% as shown in Table 11.
Thus, the result was computed using Naïve Bayes Algorithm
is an objective method that shows accuracy, precision, and
truthfulness [21]. in determining qualified faculty applicant
for teaching demonstration. Therefore, the applicant can The Level of Effectiveness of the Developed Faculty
proceed to teaching demonstration. Qualification Analysis System by The IT Experts in Terms of
(1) Functional – Correctness; (2) Efficiency/Performance
Table 12 Probability of Evidence 1 Result -Software Capacity; (3) Usability – User Error Protection; (4)
Reliability – Fault Tolerance and Recoverability; (5)
Compatibility; (6) Security; (7) Maintainability; and (8)
Portability.

IJISRT24FEB1647 www.ijisrt.com 1726

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

Table 15 Level of Effectiveness achieves mean scores of 4.0 or higher, indicating an

Characteristics Mean Verbal interpretation "Acceptable" level of performance. The overall weighted
Functional suitability 4.0 High mean of 4.4 further reinforces this assessment, positioning the
Performance efficiency 4.5 High system as acceptable across the evaluated dimensions. In
Compatibility 4.5 High summary, the system is deemed acceptable in terms of its
Usability 4.0 High functional suitability, performance efficiency, usability, and
reliability, collectively reflecting a solid performance that
Reliability 4.0 High
meets or exceeds acceptable standards.
Security 4.5 High
Maintainability 4.5 High VI. SUMMARY OF FINDINGS
Portability 4.5 High
Overall weighted mean 4.3 High After extensive research and software development, the
developed systems have the potential to change the current
Table 15 provides a comprehensive evaluation of the approach which can lead to the implementation of the system.
system's characteristics, assigning mean scores on a scale from Specifically, the recommendation process at Perpetual Help
1 to 5, where 5 represents the highest rating. The verbal College of Manila spans five working days, which is prone to
interpretation accompanying each characteristic offers errors and inefficiencies like mishandling, biases, subjective
insights into the system's performance based on these mean judgments, and conflicts due to evaluation inconsistencies.
scores. Notably, the system excels in various aspects, as
indicated by high mean scores across functional suitability, Through the use of Naïve Bayes algorithm, and machine
performance efficiency, compatibility, usability, reliability, learning from the historical data the study computed the
security, maintainability, and portability. The overall probability of a faculty applicant being qualified. The dataset
weighted mean of 4.3 underscores the system's excellence, is composed of attributes like seminars, certificates, training,
positioning it as a high-performing entity. This collective years of experience, educational background, specialization,
assessment indicates that the system effectively meets its and classification. The study identified crucial predictors,
intended functions, operates efficiently, integrates seamlessly including the number of seminars attended, certificates,
with other systems, provides a user-friendly experience, training, years of related work experience, educational
ensures reliability and security, facilitates maintainability, and background, and specialization are all instrumental in
demonstrates adaptability across different environments. In determining an applicant's qualifications.
summary, the system receives a holistic "High" verbal
interpretation, affirming its outstanding performance across In terms of effectiveness, the system achieved an overall
the evaluated characteristics and meeting or surpassing weighted mean of 4.3, accompanied by a "High" verbal
expectations. interpretation, signifying recognition of the system's
effectiveness as assessed IT experts. Furthermore, the user
The Level of User Acceptance of the Developed Faculty acceptance of the system garnered an overall weighted mean
Qualification Analysis System in Terms of the Following: (1) of 4.4, with an "Acceptable" verbal interpretation, indicating
Functional Suitability – Completeness and Appropriateness; that the system fulfilled expected functionalities with an 88%
(2) Performance Efficiency - Time-Behavior and Resource satisfaction rate.
Utilization; (3) Usability – Appropriateness, Learnability,
Operability, and User Interface Aesthetics; and Lastly (4) VII. CONCLUSIONS
Reliability – Maturity and Availability
Drawing from the findings of this study, the researcher
 The Result of the Faculty Qualification Analysis System has derived several significant conclusions about the
user Level of Acceptance is Shown in Table 11. development of the Faculty Qualification Analysis
System(FQAS). The system achieved a high verbal
Table 16 User Level of Acceptance interpretation rating, indicating its reliability in predicting
Characteristics Mean Verbal interpretation faculty classification. The inclusion of seven specific
Functional suitability 4.0 Acceptable attributes in the training dataset was found to be critical for
Performance efficiency 4.5 Acceptable accurately predicting the qualifications of faculty applicants
Usability 4.5 Acceptable for teaching demonstrations. Using collaborative data filtering
combined with the Naïve Bayes algorithm shows great
Reliability 4.5 Acceptable
potential for predicting faculty classification, thereby
Overall weighted mean 4.4 Acceptable streamlining the applicant screening process and reducing the
need for manual screening. The system demonstrated
Table 16 presents an evaluation of the system's promising results, ongoing evaluation and testing are
characteristics, with mean scores assigned on a scale ranging necessary to ensure continued success and effectiveness.
from 1 to 5, where 5 signifies the highest rating. The Overall, the FQAS holds great promise as a tool to enhance the
accompanying verbal interpretation provides insights into the accuracy and efficiency of qualification prediction processes
system's performance based on these mean scores. Across the in educational institutions.
assessed characteristics — functional suitability, performance
efficiency, usability, and reliability — the system consistently

IJISRT24FEB1647 www.ijisrt.com 1727

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

RECOMMENDATIONS [8]. Tentua, C. M., Kencana, R., & Arliansyah, A. (2020).

Job recommendation system using multi-agent system
Based on the derived conclusions, several approach. In 2020 3rd International Conference on
recommendations are proposed for potential actions. These Computer Science and Artificial Intelligence
suggestions are grounded in the study's insights to have a (ICOCSAI) (pp. 21-26). IEEE.
potential impact to bring positive outcomes and address the [9]. Mhamdi, T., Guermazi, H., & Gargouri, F. (2020). A
identified areas for improvement. recommender system for job offers based on clustering
and user behavior. In 2020 International Conference on
 Drafting of policies to govern the use and implementation Advanced Systems and Electric Technologies
of the Faculty Qualification Analysis System in (IC_ASET) (pp. 185-190). IEEE.
automating the screening process is recommended based [10]. Loor, M., & De Tré, G. (2020). Interpretable
on the result of the evaluation. predictions in naive Bayes classification using
 Testing of alternative algorithms in predicting faculty augmented appraisal degrees. Expert Systems with
classification to identify the most suitable algorithm for the Applications, 155, 113404.
Faculty Qualification Analysis System. [11]. Appadoo, S., Chavda, R., & Maharaj, T. (2020).
 Formulating a new dataset model in training the machine JobFit: A job recommendation system using machine
to enhance the system's prediction capabilities. learning techniques. In 2020 IEEE 6th International
 Implementing a feedback mechanism within the system Conference on Big Data Analytics and Computational
that would allow users to provide insights on performance Intelligence (ICBDACI) (pp. 589-594). IEEE.
and suggest areas for improvement in faculty [12]. Putra, I. K. G., Sumiati, A., & Artawa, K. (2020). A
qualifications, such as attending more seminars or Study on the Implementation of Teaching Materials in
acquiring additional certifications in specific areas. This the Pandemic Era Using ADDIE and Kirkpatrick
feedback loop can contribute to continuous refinement and Evaluation Models. Journal of Physics: Conference
optimization of the system. Series, 1567(2), 022033.
https://doi.org/10.1088/1742-6596/1567/2/022033
REFERENCES [13]. Nigam, S., Singh, S., Patel, R., & Gupta, A. (2019). A
Machine Learning Approach for Job Recommendation
[1]. Chen, S., Webb, G., Liu, L., & Ma, X. (2020). A novel using User Preferences. Proceedings of the 2019 10th
selective naïve Bayes algorithm. Knowl. Based Syst., International Conference on Computing,
192,105361. https://doi.org/10.1016/j.knosys.2019.10 Communication and Networking Technologies
5361. (ICCCNT), 1-6. doi:
[2]. Pradeep Kumar Singh, Pijush Kanti Dutta Pramanik, 10.1109/icccnt45670.2019.8945067.
Avick Kumar Dey,Prasenjit Choudhury (2021). [14]. Talukdar, D., Saha, S., & Hazarika, S. M. (2018).
Recommender Systems: An Overview, Research Analysis of Named Entity Recognition system using
Trends and Future Directions. International Journal of different feature sets. Journal of King Saud
Business and Systems Research 15(1): 14-52, DOI: University-Computer and Information Sciences, 30(1),
10.1504/IJBSR.2021.10033303 1-10. doi: 10.1016/j.jksuci.2016.10.001
[3]. Mohanty, Sachi Nandan; Chatterjee, Jyotir Moy; Jain [15]. Sugahara, K., Hirasawa, K., & Yamanishi, K. (2018).
Sarika; Elngar, Ahmed A. ; Guta, Priya (2020). Exact and Approximate Learning in Bayesian
Recommender System with Machine Learning and Networks: Comparing the Performance of Conditional
Artificial Intelligence. Doi:10. 1002/9781119711582 Log Likelihood and Marginal Likelihood. IEEE/ACM
[4]. Kaviani, A., & Dhotre, P. (2017). Naive Bayes Transactions on Computational Biology and
Classifier for Text Classification with Python. Bioinformatics, 15(1), 267-275.
International Journal of Computer Applications, https://doi.org/10.1109/TCBB.2017.2650818
167(10), 37-40. DOI: 10.5120/ijca2017914383. [16]. Wei, J. (2018). Naive Bayes Classifier. Towards Data
[5]. Autoridad Nacional del Servicio Civil. (2021). Science. Retrieved from
Recommendation system for employment in https://towardsdatascience.com/naive-bayes-classifier-
career-based social networking websites. 81d512f50a7c
[6]. Tayade, S. B., Shinde, S. V., & Patel, R. B. (2021). A [17]. Urera, F. L. J. & Balahadia, F. F. (2019).
review of job recommender systems. International ICTeachMUPO: An evaluation of information
Journal of Computer Science and Mobile Computing, e-learning module system for faculty and students.
10(3), 232-240. International Journal of Computing Sciences Research,
[7]. Puspasari, A., Irwansyah, A., & Nurhadiyanto, P. 3(1), 163-188. doi: 10.25147/ijcsr.2017.001.1.31.
(2021). Job vacancy recommendation system for [18]. Tentuaa, M. N., & Supraptob. (2020).
undergraduate students using the K-Means clustering Recommendation System Using Machine Learning: A
method. Journal of Physics: Conference Series, Review Paper. International Journal of Innovation,
1827(1), 012062. Creativity and Change, 12(11), 40. Retrieved from
www.ijicc.net.

IJISRT24FEB1647 www.ijisrt.com 1728

Volume 9, Issue 2, February 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24FEB1647

[19]. Valenzuela, Alberto, Recruitment and Selection

Process of Faculty in Higher Education Institutions in
the Philippines (August 30, 2019). Available at SSRN:
https://ssrn.com/abstract=3445566 or
http://dx.doi.org/10.2139/ssrn.3445566
[20]. Dayal, A. (2015). Recruitment and selection process
(PGDBA Thesis, HR). Symbiosis Center For Distance
Learning, Pune. (Publication No. 78)
[21]. Kapur, R. (2018, March). Recruitment and Selection.
Research Gate.
[22]. Tomlinson, G., & Freeman, S. (2017). Who really
selected you? Insights into faculty selection processes
in top-ranked higher education graduate
programmes. Journal of Further and Higher
Education, 42, 855 -
867. https://doi.org/10.1080/0309877X.2017.1323192
[23]. Upadhyay, A., & Khandelwal, K. (2018). Applying
artificial intelligence: implications for recruitment.
Strategic HR Review.
https://doi.org/10.1108/SHR-07-2018-0051.
[24]. Nawaz, N. (2019). Artificial Intelligence Interchange
Human Intervention in the Recruitment Process in
Indian Software Industry. Labor: Personnel Economics
eJournal. https://doi.org/10.2139/ssrn.3521912.
[25]. Kumar, C., Kumar, L., Muthyalappa, C., Shruthi, D.,
Shareena, S., & Sireesha, D. (2021). An Integrated
College Faculty Recruitment System. International
Journal of Scientific Research in Science and
Technology. https://doi.org/10.32628/ijsrst2183140.
[26]. Singh, J., Salvi, R., Surve, J., & Sawant, A. (2023). A
Comprehensive Campus Recruitment and Placement
System for Optimizing the Hiring
Process. International Journal for Research in Applied
Science and Engineering Technology.
https://doi.org/10.22214/ijraset.2023.50800.
[27]. Tatsat, H., Puri, S., & Lookabaugh, B. (n.d.). Machine
learning and data Science blueprints for finance.
O’Reilly Online Learning.
https://www.oreilly.com/library/view/machine-learnin
g-and/9781492073048/ch04.html

IJISRT24FEB1647 www.ijisrt.com 1729