Anodot Enhances Machine Learning Performance with Intel Solutions

This case study details how Anodot improved its machine learning anomaly detection and forecasting capabilities by leveraging Intel® Xeon® Scalable Processors and Intel® software tools. The document outlines the challenges faced by Anodot, including the need for unlimited scalability and effective management of compute costs as their customer base grew. It describes the specific machine learning models used, such as the Autocorrelation Function (ACF) for anomaly detection and XGBoost for forecasting, and highlights the performance improvements achieved through optimization. The results indicate that Anodot experienced a training performance increase of up to 127 times for the ACF algorithm and a fourfold increase in inference speed for the XGBoost algorithm. The case study emphasizes the significance of these enhancements in enabling Anodot to provide real-time monitoring and forecasting services efficiently. Additionally, it mentions the components of the solution, including Intel® oneAPI Data Analytics Library and Intel® IPP, which contributed to the overall performance gains.