A framework of python to do general machie learning using sk-learn, numpy, matplotlib and pandas
Help you prepare data, train and save models, handle unbalanced problem, plot decision margin,
save tree into file, give you feature importance and top possible label. Validate is precision,
recall, ROC and F-measure,etc
Now is only suit for Classification, later will update
File introduction
TrainModels.py the main file if you want to train or test models
MLUtils.py some tools for machine learning use
ModelZoo.py models is here, such as SVM,GBDT
DataPrepare.py prepare your data before training
User guide
install python2.7, better is Anaconda
modify DataPrepare.py and set your data path before training
modify TrainModels.py and set your path of model and log, then choose some model to train
If you think this repo helps you, may you give me a star :)