These programs learn motions, evaluate results, rank what datas is nearest in target data.

##Preparation

I use datas in UTD-MHAD

##Installation

pip install -r requirements.txt

And,python version is 2.7.9 .

if your machine is MAC, create matplotlibrc in ~/.matplotlib and wirte below to write graph.

backend : TkAgg

##Details

• learning.py :

To learn motions. It saves results (default: 'test.state' and 'test.model').

usage: learningData.py [-h] [--dir DATA_DIR] [--save SAVE_FILENAME]
optional arguments:
 -h, --help            show this help message and exit
 --dir DATA_DIR
 --save SAVE_FILENAME
 

If you want to change data, change i_data in learning.py

• evaluate.py :
  Evaluate results of learning.py. It uses accuracy, f1-score, auc, hamming loss ,and make confusion matrix.

• ranking.py : rank what datas is nearest in target data.