CMPT353 PROJECT : BODY ACTIVITY MONITORING

In this project, we predict a user's activity based on whether they are standing, walking or running.

A Report with all the analysis and details of the used techniques is saved as the file CMPT 353 Report.pdf

Table of contents

• General info
• Technologies
• Libraries Used
• Features
• Order Of Execution
• Status
• Based On
• Created By

General info

The idea behind this project is to predict based on a user’s activity whether they are standing, walking, or running using machine learning algorithms. Data was collected through sensors and data preprocessing was done. Statistical tests were used to analyze the data. Feature engineering was used for training Machine Learning Models.

Technologies

You will need to install Anaconda

• Python - version 3.0

Libraries Used

• numpy
• pandas
• matplotlib
• statsmodels
• scipy
• sklearn
• joblib

Features

List of features ready

• Data preprocessing
• Statistics
• Machine Learning

Order Of Execution

     1) Run '01-Analysis.ipynb'
        - Results produced: filtered files for each scenario
  
            
    2) Run '02-Statistics.ipynb'
        - Results produced: one transformed file and graphs of multiple inferential and statistical tests
          
        
    3) Run '03-MachineLearning.ipynb'
        - Results produced:Classification report, ROC Curve and Confusion matrix for each model 
        - Models saved in location: "Models"
        
    4) Run '04-Prediction.ipynb'
       - Imports Saved models from "Models" 
       - Predicts and print results for never seen data in "Data/testData"
       
    5) Run main.py on terminal
       -Command will look like:
       -python3 main.py data/testdata/walk3.csv
       -Results shown on teminal

Status

Project is: finished

Based on

Project based on Computational Data-Science

Created By

Created by Arpit Kaur,Sharjeel Ahmad and Srimalaya Ladha