InterpreTS 0.5.0

Feature extraction from time series to support the creation of interpretable and explainable predictive models.

InterpreTS - Overview

interpreTS is a Python library designed for extracting meaningful and interpretable features from time series data to support the creation of interpretable and explainable predictive models.

Key Features

• Feature Extraction: Extract features such as mean, variance, spikeness, entropy, trend strength, and more.
• Interpretable Models: Generate explainable predictive models by leveraging extracted features.
• Streaming Data Support: Process and extract features in real-time from streaming data sources.
• Scalability: Supports parallel and distributed computation with joblib and dask.
• Custom Features: Extend the library with user-defined features.
• Validation: Ensures input data meets the required format and quality using built-in validators.

Requirements

• Python 3.10 or above
• pandas==2.2.3
• numpy
• statsmodels==0.14.4
• langchain_community==0.3.14
• langchain==0.3.14
• openai==1.59.4
• streamlit==1.41.1
• scikit-learn
• joblib==1.4.2
• tqdm==4.67.1
• dask==2024.12.1
• scipy==1.15.0
• pillow==11.1.0

Installation Guide

Follow these steps to install InterpreTS and its dependencies:

From PyPI

pip install interpreTS

From Source

1. Clone the repository:

git clone https://github.com/yourusername/interpreTS.git
cd interpreTS

2. Install dependencies: Install the required packages listed in the requirements.txt file:

pip install -r requirements.txt

3. Install InterpreTS: Run the following command to install InterpreTS:

pip install .

GUI

InterpreTS offers a user-friendly GUI to facilitate time series feature extraction and model explainability.

Accessing the GUI

The GUI is hosted online and available at InterpreTS GUI.

Alternatively, you can run it locally:

1. Navigate to the GUI directory:

cd interpreTS/core/gui

2. Launch the GUI using Streamlit:

streamlit run gui.py

Features of the GUI

• Feature Extraction: Easily upload time series data and extract predefined or custom features.
• Visualization: View extracted features directly in the GUI.

Verifying Installation - Example: Basic Feature Extraction

Once installed, you can verify the installation by running a simple feature extraction example:

import pandas as pd
from interpreTS import FeatureExtractor, Features

# Sample time series data
data = pd.DataFrame({
"time": pd.date_range("2023-01-01", periods=10, freq="D"),
"value": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
})

# Initialize the FeatureExtractor
extractor = FeatureExtractor(features=[Features.MEAN, Features.VARIANCE], feature_column="value")

# Extract features
features_df = extractor.extract_features(data)
print(features_df)

Additional Usage Example with Time Series Data

import pandas as pd
import numpy as np
import time
from interpreTS import FeatureExtractor, Features

def report_progress(progress):
   print(f"Progress: {progress}%", flush=True)

# Generate synthetic time series data
data = pd.DataFrame({
'id': np.repeat(range(100), 100),  
'time': np.tile(range(100), 100),  
'value': np.random.randn(10000)  
})

# Initialize the FeatureExtractor
feature_extractor = FeatureExtractor(
features=[Features.ENTROPY],
feature_column="value",
id_column="id",
window_size=5,  # Rolling window size
stride=2        # Step size for moving the window
)

# Measure execution time
start_time = time.time()

# Extract features
features_df = feature_extractor.extract_features(data, progress_callback=report_progress, mode='sequential')

end_time = time.time()

# Display results and execution time
print(features_df.head())  # Display the first few rows of the resulting DataFrame
print(f"Execution time: {end_time - start_time:.2f} seconds")

Documentation

Complete documentation is available on GitHub Pages.

Issues and Support

For any issues, please consult our Issue Tracker on GitHub.