Eagle-AI-GUI

GUI for Machine-Learning Algorithms

     

TOOLS/LANGUAGES USED

GUI is made with tkinter a python based GUI framework see tutorial here

ML algorithms are coded from scratch based on lectures of adrew ng. course see course here

Installation

Step-1: Download and extract the files

Step-2: Create a virtual enviroment(optional)

Step-3: open the path where files are extracted in terminal

Step-4: type pip install -r requirements.txt and hit enter in terminal

Step-5: type python main.py and hit enter to start the application

Note->In windows application may show not responding, just leave it idle for few seconds

Usage Guide

• Linear Regression

  Sample file is in samples/Linear Regression directory

  Step-1: This is first screen that will appear when the program is executed, load in your Training set, DEV set and Test set and press initialize.

  

  Note->

  • You must load in only .csv, .xls or .xlsx file
  • Columns of the file uploaded will be used as parameters(features)
  • Column of one is added by the application and it should not be present in original file
  • It is considered a good practice to have different test and Dev(cross validation) set but same file can be uploaded in both
  • Normalization is applied automatically by the application

  Step-2: After Initialization this screen will appear, from the dropdown list choose "Linear Regression" fill in the hyper parameters than press START TRAINING.

  

  Progress Bar below "START TRAINING" button denotes the progress in model training.

  Step-3: After the training is finished a pop-up message will appear stating the successfull completion of model training.

  To save the model press SAVE MODEL this will export three files one containing parameters and other two containing normalization factors

  

• Logistic Regression

  Sample file is in samples/Logistic Regression directory

  Step-1: Same as in Linear Regression

  Note->

  • You must load in only .csv, .xls or .xlsx file
  • Columns of the file uploaded will be used as parameters(features)
  • Column of one is added by the application and it should not be present in original file
  • It is considered a good practice to have different test and Dev(cross validation) set but same file can be uploaded in both
  • Normalization is applied automatically by the application
  • Multi-Class classification is not supported(currently) so y column should have only zeros or one

  Step-2: After Initialization this screen will appear, from the dropdown list choose "Logistic Regression" fill in the hyper parameters than press START TRAINING.

  

  Step-3: After the training is finished a pop-up message will appear stating the successfull completion of model training.

  To save the model press SAVE MODEL this will export three files one containing parameters and other two containing normalization factors

  

• Deep Neural Network Sample file is in samples/Deep Neural Network directory

  Step-1: Same as in Linear Regression

  Note->

  • You must load in only .csv, .xls or .xlsx file
  • Columns of the file uploaded will be used as parameters(features)
  • Column of one is added by the application and it should not be present in original file
  • It is considered a good practice to have different test and Dev(cross validation) set but same file can be uploaded in both
  • Normalization is applied automatically by the application
  • Multi-Class classification is not supported(currently) so y column should have only zeros or one
  • Currently only a three layered neural network is supported

  Step-2: After Initialization this screen will appear, from the dropdown list choose "Logistic Regression" fill in the hyper parameters than press START TRAINING.

  Note->

  • optimizer takes in value gd for Gradient Descent, momentum for GD+momentum and adam for Adam optimization.
  • hyper parameter Beta is for momentum optimization with default value=0.9(leave it empty for other optimizations)
  • hyper parameters Beta1 and Beta2 are for Adam optimization with default value=0.9,0.999(leave it empty for other optimizations)
  • example for layer input 5,2,1 last layer(output layer) should be 1 and each layer should be seperated by comma
  

  Step-3: After the training is finished a pop-up message will appear stating the successfull completion of model training.

  To save the model press SAVE MODEL this will export three files one containing parameters and other two containing normalization factors

  

  Contributions

  1.Fork it!

  2.Clone the forked repository to local system.

  3.Read issues and solve it, or add your own issues 😊

  4.Commit your changes: git commit -m 'Add some feature'.

  5.Push to the feature branch

  7.Submit a pull request 😄

If you had fun, consider to give a star ⭐ to this repository