Learn to use Non-Transformed Temporal Similarity Search (Non-Transformed TSS), a model designed to identify patterns and trends directly upon incoming fast flowing time series data. Non-Transformed TSS acts on time series columns directly with no need to embed or build an index to execute vector search. This method is a great fit to handle flowing data where there is not time to build an index and similarity search must be run quickly.
In this sample, we will use KDB.AI to perform Non-Transformed TSS over example time series market data. The sample will dive into temporal vector searches using Non-Transformed TSS, as well as hybrid searches over this market data.
Download the Jupyter Notebook at the GitHub repository.