Repository containing notebooks of my posts on MEDIUM.
To be notified every time a new post is published, SUBSCRIBE HERE.
- Proxy SHAP: Speed Up Explainability with Simpler Models [post][code]
- Time Series Forecasting in the Age of GenAI: Make Gradient Boosting Behaves like LLMs [post][code]
- Hitchhiker’s Guide to MLOps for Time Series Forecasting with Sklearn [post]|[code]
- Hitting Time Forecasting: The Other Way for Time Series Probabilistic Forecasting [post]|[code]
- Forecasting with Granger Causality: Checking for Time Series Spurious Correlations [post]|[code]
- Hacking Causal Inference: Synthetic Control with ML approaches [post]|[code]
- Model Selection with Imbalance Data: Only AUC may Not Save you [post]|[code]
- PCA for Multivariate Time Series: Forecasting Dynamic High-Dimensional Data [post]|[code]
- Hacking Statistical Significance: Hypothesis Testing with ML Approaches [post]|[code]
- Time Series Forecasting with Conformal Prediction Intervals: Scikit-Learn is All you Need [post]|[code]
- Rethinking Survival Analysis: How to Make your Model Produce Survival Curves [post]|[code]
- Extreme Churn Prediction: Forecasting Without Features [post]|[code]
- Forecast Time Series with Missing Values: Beyond Linear Interpolation [post]|[code]
- Forecasting Uncertainty with Linear Models like in Deep Learning [post]|[code]
- Time Series Forecasting with Feature Selection: Why you may need it [post]|[code]
- Anomaly Detection in Multivariate Time Series with Network Graphs [post]|[code]
- How to Improve Recursive Time Series Forecasting [post]|[code]
- Retrain, or not Retrain? Online Machine Learning with Gradient Boosting [post]|[code]
- Data Drift Explainability: Interpretable Shift Detection with NannyML [post]|[code]
- Word2Vec with Time Series: A Transfer Learning Approach [post]|[code]
- SHAP for Drift Detection: Effective Data Shift Monitoring [post]|[code]
- Forecasting with Trees: Hybrid Classifiers for Time Series [post]|[code]
- Boruta SHAP for Temporal Feature Selection [post]|[code]
- Forecasting with Trees: Hybrid Modeling for Time Series [post]|[code]
- Recursive Feature Selection: Addition or Elimination? [post]|[code]
- Improve Random Forest with Linear Models [post]|[code]
- Is Gradient Boosting good as Prophet for Time Series Forecasting? [post]|[code]
- Linear Boosting with Automated Features Engineering [post]|[code]
- Improve Linear Regression for Time Series Forecasting [post]|[code]
- Boruta and SHAP for better Feature Selection [post]|[code]
- Explainable AI with Linear Trees [post]|[code]
- SHAP for Feature Selection and HyperParameter Tuning [post]|[code]
- Model Tree: handle Data Shifts mixing Linear Model and Decision Tree [post]|[code]
- Add Prediction Intervals to your Forecasting Model [post]|[code]
- Linear Tree: the perfect mix of Linear Model and Decision Tree [post]
- ARIMA for Classification with Soft Labels [post]|[code]
- Advanced Permutation Importance to Explain Predictions [post]|[code]
- Time Series Bootstrap in the age of Deep Learning [post]|[code]
- Anomaly Detection with Extreme Value Analysis [post]|[code]
- Time Series generation with VAE LSTM [post]|[code]
- Extreme Event Time Series Preprocessing [post]|[code]
- One-Class Neural Network in Keras [post]|[code]
- Real-Time Time Series Anomaly Detection [post]|[code]
- Entropy Application in the Stock Market [post]|[code]
- Time Series Smoothing for better Forecasting [post]|[code]
- Time Series Smoothing for better Clustering [post]|[code]
- Predictive Maintenance with ResNet [post]|[code]
- Neural Networks Ensemble [post]|[code]
- Anomaly Detection in Multivariate Time Series with VAR [post]|[code]
- Corr2Vec: a WaveNet architecture for Feature Engineering in Financial Market [post]|[code]
- Siamese and Dual BERT for Multi Text Classification [post]|[code]
- Time Series Forecasting with Graph Convolutional Neural Network [post]|[code]
- Neural Network Calibration with Keras [post]|[code]
- Combine LSTM and VAR for Multivariate Time Series Forecasting [post]|[code]
- Feature Importance with Time Series and Recurrent Neural Network [post]|[code]
- Group2Vec for Advance Categorical Encoding [post]|[code]
- Survival Analysis with Deep Learning in Keras [post]|[code]
- Survival Analysis with LightGBM plus Poisson Regression [post]|[code]
- Predictive Maintenance: detect Faults from Sensors with CRNN and Spectrograms [post]|[code]
- Multi-Sample Dropout in Keras [post]|[code]
- When your Neural Net doesn’t know: a bayesian approach with Keras [post]|[code]
- Dynamic Meta Embeddings in Keras [post]|[code]
- Predictive Maintenance with LSTM Siamese Network [post]|[code]
- Text Data Augmentation makes your model stronger [post]|[code]
- Anomaly Detection with Permutation Undersampling and Time Dependency [post]|[code]
- Time2Vec for Time Series features encoding [post]|[code]
- Automate Data Cleaning with Unsupervised Learning [post]|[code]
- People Tracking with Machine Learning [post]|[code]
- Time Series Clustering and Dimensionality Reduction [post]|[code]
- Anomaly Detection in Images [post]|[code]
- Feature Importance with Neural Network [post]|[code]
- Anomaly Detection with LSTM in Keras [post]|[code]
- Dress Segmentation with Autoencoder in Keras [post]|[code]
- Extreme Event Forecasting with LSTM Autoencoders [post]|[code]
- Zalando Dress Recommendation and Tagging [post]|[code]
- Remaining Life Estimation with Keras [post]|[code]
- Quality Control with Machine Learning [post]|[code]
- Predictive Maintenance: detect Faults from Sensors with CNN [post]|[code]