RAPID machine learning Time Series Analysis python interface specifications¶
This document describes specifications of the RAPID machine learning Time Series Analysis(RAPID TSA) python interface.
Overview¶
This software is RAPID TSA python interface, which is python interface for RAPID TSA V2.2.
Below is procedure with API of RAPID TSA python interface.
Note
The procedure is the same with RAPID TSA V2.2.
For train
- Calculates the statistics
- Pre-processes the train data
- Trains with the label data, and generate the model
For predict
- Pre-processes the predict data with train’s statistics
- Predicts the label with the model
Quickstart¶
import os
ROOT_DIR = os.path.abspath(os.path.curdir)
DATA_DIR = os.path.join(ROOT_DIR, 'data')
# execute preprocess for train
from rapid_tsa_python import exec_preprocess
train_label_path = os.path.join(DATA_DIR, 'train_label_cls.lab')
preprocess_def_path = os.path.join(DATA_DIR, 'preprocess_cls.json')
exec_preprocess(preprocess_def_path, label_path=train_label_path)
# execute train
from rapid_tsa_python import exec_train
train_preprocessed_label_path = os.path.join(DATA_DIR, 'train_label_cls_preprocessed.lab')
train_parameter_conf_path = os.path.join(DATA_DIR, 'train_param.conf')
model_dir = os.path.join(ROOT_DIR, 'model')
exec_train('cls', '1DCNN', train_preprocessed_label_path, model_dir,
param_conf_path=train_parameter_conf_path)
# execute preprocess for predict
predict_label_path = os.path.join(DATA_DIR, 'predict_label_cls.lab')
exec_preprocess(preprocess_def_path, label_path=predict_label_path)
# execute predict
from rapid_tsa_python import exec_predict
predict_preprocessed_label_path = os.path.join(DATA_DIR, 'predict_label_cls_preprocessed.lab')
result_dir = os.path.join(ROOT_DIR, 'result')
exec_predict(model_dir, label_path=predict_preprocessed_label_path, output_dir=result_dir)
Input Specification¶
Data and config files of RAPID TSA python interface are same with RAPID TSA V2.2. So, refer to chapter 6 of RAPID TSA V2.2 User Guide.
Others¶
Logging and Error messages¶
Logging and error messages are output in log file following with RAPID TSA V2.2. So, refer to chapter 6.7 of RAPID TSA V2.2 User Guide.
Note
You can see log file with cat, less or etc. command at Jupyter Terminal. And log file is outputted at /var/log/nec/rapid/TSA_<index>.log and /var/log/nec/rapid/MLE_<index>.log by default.