Detection of switching time
In this thesis, we talk about detecting switches and jumps in the data. Every dataset has a certain hidden pattern in it. These patterns are sometimes irregular and abnormal because of the jumps and switches in the datasets or sometimes because of outliers in the datasets. At the point of out lier, data changes it's pattern and it looks irregular. Generally people use control limiting to detect those changes or outliers, but using control limiting one can not detect the minor changes. So, we use change point analysis along with splines and Gibbs phenomenon to detect minor shifts or changes in the dataset. But, we can not use change point analysis in stead of control charting, both methods are used in complementary fashion. Control charting is useful to detect single outliers, and change point analysis detects bunch of outliers at a time. Here our dataset is a mixture of two data sets, in which one dataset is a set of outliers. We are trying to separate these two datasets using change point analysis, and trying to find the change point in between these two datasets, This change point is called the shift or jump.