matlab预测电池寿命程序代码-Cycle-life-prediction-using-machine-learning:循环...

# Cycle-life-prediction-using-machine-learning This research was based on the work of students at Stanford university titled “Data-driven prediction of battery cycle life before capacity degradation,”. They had created a dataset, the largest open source of its kind, and had used machine learning to predict lithium ion battery life. The aim of my research was to first recreate their data, and then finally create my own model to rival the accuracies of this project using the same dataset. The datasets used in this study are available at https://data.matr.io/1. ## results_recreation.m Purpose: Loads the three batches of data and combines it into one large dataset on matlab. Alters some of the innacurate values for the cycle life. The code then extracts and manipulates the relevant data to create the csv file required to run the Elastic net models. Requires: Matlab, the three datasets Typical runtime is a few minutes ## variance_data.csv Purpose: The csv file containing the variance data alongside the cycle life for all 124 cells. This file is altered slightly by providing headings for each column. This is needed when running the python program. Requires: None ## Data_recreation.ipynb Purpose: Generates the Elastic net for the variance, cycle life dataset. This code calls the csvfile into the dataset, and prepares the data to be put into the Elastic net. The data is split into training, test, and secondary test according to the same distribution as the Stanford paper. The code then displays the results graphically and prints the three values for the mean percent error of the training, test and secondary test data. Requires: This code requires the use of the numpy, pandas, matplotlib.pyplot and sklearn.linear_model libraries. Previous download of the variance_data.csv file is required to run this program. Typical runtime is a few seconds.