Temporal causal modeling algorithms

Abstract

This paper talks about Granger Causality Test and Temporal Causal Modeling(TCM) in IBM SPSS Modeler. It refers to time series concepts, data mining knowledge and statistical concepts: such as autoregression, linear regression and F-test. It also shows how we can use TCM to build a good model, and how to read Granger Causality graphs(in SPSS modeler it is called impact graph). And how to judge if the model is good or not, there are many indexes and concepts used to judge. Also we will show how to use python to apply granger causality tests. After showing the indexes and the concepts, we will use granger causality and TCM to analyze some real question, try to find out if humidity has an effect on temperature. In the end, we will list some real analysis in other areas such as the economy using granger causality.