ML for anomalous diffusion model

ML for anomalous diffusion modelDapeng Wang7.24.2025This repository contains the necessary codes written in Python 3 to train the classifiers.1. Required modules1. numpy, pandas, scipy, scikit-learn, matplotlib and seaborn for basic scientific computing / data processing.2. hmmlearn for trajectory / diffusion feature extraction.3. tqdm and natsort for progress bar / file traversal.4. joblib for storing the classifiers.Usage1.Download the whole repository and extract it in a directory of your choice.2.Use the Trajs_simulation_with_noise.nb file to generate training data.3.Use the features_zyh_ATTM.py, features_zyh_BW.py, features_zyh_CTRW.py, features_zyh_FBM.py, features_zyh_LW.py, and features_zyh_SBM.py to compute the most discriminative feature vectors for each diffusion model to support subsequent classifier training and prediction.4.Use the findhypara_step2.py file to use RandomizedSearchCV for hyperparameter tuning of a RandomForestClassifier and save the results to a JSON file.5.Use the model_step3.py file to perform classification prediction using a RandomForestClassifier.6.Use the features_exp_step4.py files to load trajectory data, calculate features for each trajectory, and save the results as a CSV file.7.Use the model_predict_step5.py code to load a pre-trained machine learning model, make predictions on feature data, analyze prediction results, and save outputs as CSV files.