Kill-Nickson / sport_predicting_ffnn

A sport predicting software based on using a feed-forward neural network (created from scratch) is written using python. This repo is meant for demonstration purposes only.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

sport_predicting_ffnn


Project's modules:

  • FFNN - a module, which contains class of a feed forward neural network. The n_network is a reproduced and reworked version of a network shown there - (source: https://github.com/makeyourownneuralnetwork/makeyourownneuralnetwork/blob/master/part3_neural_network_mnist_data_with_rotations.ipynb). The changes made the network more universal because now it allows to create any amount of its layers and nodes in each layer.
  • weights_recorder - a module, which allows to init, learn and train the FFNN to save weights of the network, in order reuse them later.
  • opt_live_predictor - its main purpose is to parse a sport statistics web-site and return a list of data, required for predicting a sport event results (contains only a template of how it might look).
  • gui - a module, which contains several classes, that are responsible for predicting sport events and visualization of the predictions.

UI-cases:

For running all possible states of the program interface:

  1. Comment out line 26 in the gui.py module: comment
  2. Uncomment any out of 5 commented test cases placed in lines from 29th to 54th in the gui.py module: case
  3. Run gui.py.

If you made everything right, you would be able to see similar to on of the next two screenshots: gui_cases

About

A sport predicting software based on using a feed-forward neural network (created from scratch) is written using python. This repo is meant for demonstration purposes only.


Languages

Language:Python 100.0%