Hidden-Markov-Model
- scipy>=0.15.1
- numpy>=1.8.0
- main.py: primary file to run. Contains implementation to parse input and call inference and learning modules.
- hmms.py: Implementation of forward backward functions, inference and learning methods.
- data.txt: Observable outcome for time series data. total 50 instance, each of length 30.
- model-input-file.txt: Pre-estimated parameters for inference.
-
Perform Inference: Infer hidden status given outcomes of time series data.
python main.py --mode test --data data.txt --predictions-file predictions-file.txt \ --model-input-file model-input-file.txt -
Learning-Estimate model Parameters:
python main.py --mode train --data data.txt --model-input-file model-input-file.txt \ --model-output-file model-output-file.txt --iterations 100 -
Predictions: Predict the hidden states for new data using the parameter learned.
python main.py --mode test --data data.txt --predictions-file predictions-file-after-learning.txt \ --model-input-file model-output-file.txt\ -
For more details run help:
$ python main.py --help usage: main.py [-h] --mode {train,test} [--data DATA] [--predictions-file PREDICTIONS_FILE] [--model-input-file MODEL_INPUT_FILE] [--model-output-file MODEL_OUTPUT_FILE] [--iterations ITERATIONS] Perform HMM inference or learning.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent