yujianyuanhaha / BerPredict

Bit Error Rate Predictor using Neural Network

Bit-Error-Rate (BER) Predict using Neural Network

Aug, 2019
Jet Yu, [e-mail]
Hussein Metwaly Saad, [e-mail]
Yue Xu, [e-mail]
Mike Buehrer, [e-mail]
Wireless, ECE, Virginia Tech

Dataset

under folder DATA, for more data, click DataSet, and ./DATA/READ.md describe the data.

How to run

• if need, run ./FUNCTION/mainData.m to generate dataset
• Neural Network
  • main_regression.py, where labels are in [0,1] scale
  • main_regression_dB.py, where labels are in log scale, [0, 40]dB
  • main_classification.py, where labels are in another log scale
• Random Forest
  • mainRF.m, in matlab. 1-dimension
  • mainRF.py, in python

File Description

• ToolBox.py is called by main_classification.py.
• Under FUNCTIONS run mainGenerateData.m to generate traing dataset.
• Result folder has all trained results when run main*.py file
• Predict folder has all test results when run test.py

Chat/ Hangout Group

Chat/ Hangout Group

Slides

performance Google slides

News

(Sep 29) starter code online

Roadmap

Config

running on ARC VT is prefered, run python is preferred.

• Python
  • tensorflow, Keras, scikitlearn

Running Time Reference