Machine learning with ClickHouse and DoubleCloud

The article discusses the integration of machine learning with ClickHouse, an open-source column-oriented database management system. It highlights that ClickHouse has built-in functions for stochasticLinearRegression and stochasticLogisticRegression, as well as the ability to evaluate against external CatBoost models using catboostEvaluate function. The author also introduces MindsDB, an open-source framework that provides a familiar SQL interface for in-database SQL-based machine learning. It is based on MySQL syntax and can be used with various open-source machine learning frameworks. The article demonstrates how to implement MindsDB with DoubleCloud, a managed ClickHouse DBMS service, by creating a log table, importing data into it, converting the table to MergeTree, and then connecting MindsDB to ClickHouse. Finally, the author explains how to train models in MindsDB and test their accuracy using sample queries.