Skip to content
Training and executing MLPs with hidden layers
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
README.md
executeDT.py
executeMLP.py
test_data.csv
trainDT.py
trainMLP.py
train_data.csv
weights0.csv
weights10.csv
weights100.csv
weights1000.csv
weights10000.csv

README.md

Multi-Layer-Perceptron

Training and executing MLPs with hidden layers
Files in this directory:

  1. trainMLP.py: Program to train the MLP using train_data.csv file. Takes commad line arugment. 'Usage: python trainMLP.py samplefile'

  2. executeMLP.py: Executes the trained MLP on test data using test_data.csv. 'Usage: python executeMLP.py weights.csv sample'

  3. trainDT.py: This program takes the training data file as input via the command line.
    This program display the min.max,avg depth, number of nodes and leaves before and after pruning. To run this porgram - python3 trainDT.py train_data.csv
    Multiple plots are generated after every split.

  4. executeDT.py: This program takes the testing data file as input and provides the classification of data as at plot graphs.
    This program also calculates confusion matrix and total profit before and after pruning.
    To run this program - python3 trainDT.py test_data.csv

  5. train_data.csv: Training data file

  6. test_data.csv: Testing data file

  7. weights0.csv: weight values at initial stage

  8. weights10.csv: weight values after 10 Epocs

  9. weights100.csv: weight values after 100 Epocs

  10. weights1000.csv: weight values after 1000 Epocs

  11. weights10000.csv: Weight values after 10000 Epocs

    How to run the program files:

    1. Make sure that all program and date files are in same folder
    2. Run python programs by following how to run information given along with that.
    3. A graph will be plotted which shows specific epoc vs sum of squared error for samples in each epoc for trainMLP.py. weight files are created in the same folder based on the Epoc values.
    4. After running executeMLP.py program, results will be displayed in console along with a graph where all points are plotted with different colorrs or shape based on thier classes.
You can’t perform that action at this time.