Having used the SAP HANA Predictive Analysis Library (PAL) for a few years now, I had to post a short video to demonstrate the vast improvement made in SAP HANA SPS 09.
This example uses linear regression (LR) to predict stock movements based on the movement of another closely correlated stock and the S&P 500 index. The goal is to just show a simple LR example that is easily understood, nothing more. Having the historical data we can train the algorithm and make our forecast all in one flowgraph model, procedure. We can then check the results with our test data. Since the model uses historical data it is not useful for predicting future stock movements, but shows how you could guess the movement of one stock if you only had the corresponding data of a closely correlated stock and the S&P500.
For further details, visit the HANA Academy videos on the SAP HANA Application Function Modeler (AFM) found here.
