Method
ARIMA / VAR
Classical autoregressive and differencing models with stationarity tests.
7 cases from 5 companies apply ARIMA / VAR. 3 report results in production.
Typical tasks: Detect anomalies, Forecast demand and load, Reduce ML cost, Explain a metric change. Companies: Lyft, T-Bank, VTB, trivago and 1 more. Often combined with: Anomaly detection, Ensembling, Prophet, Knowledge distillation, Linear regression and 16 more.
| Virtual Sensors in Industry | manufacturing | 2025-05 | |
| Real-Time Spatial Temporal Forecasting @ Lyft | mobility | 2025-05 | |
Foundation Models: Is There Life in Time Series?![]() | ![]() | general | 2024-05 |
| Anomaly Detection: Anomaly Scoring for Preventive Detection of Information System Failures | banking | 2024-05 | |
| Real-world Insights: Anomaly Detection in Internet Traffic | travel | 2024-02 | |
| ML for Forecasting and Analyzing Passenger Transport Behavior | — | mobility | 2023-05 |
| Building a large scale unsupervised model anomaly detection system — Part 2 | mobility | 2023-04 |
