/two-layer-n-n

Two layer neural network for classification and regression problem

Primary LanguagePythonMIT LicenseMIT

How to run

  • pip install -r requirements.txt
  • python concrete.py

Specification

  • This code has one hidden layer with 50 neurons and iterates 1000 times to find the regression mean squared error and classification accuracy in case of regression and classification respectively.

  • It also outputs regression prediction and class in case of regression and classification problem respectively.

  • It also outputs the ROC curve for classification problem.

  • It also outputs the scatter plot for regression problem.

Note : Code takes time to run, therefore added the screenshot of output and plots

Scatter Plot for regression

alt text

ROC Curve for classification

alt text

Screenshot of output

alt text