In various situations, such as natural phenomena and production activities, the state of things can change at a certain point in time. In such cases, it's crucial to identify these changes from data and pinpoint the underlying causes. However, when the number of variables increases, it becomes difficult to grasp the overall picture, including the relationships between the variables, making it challenging to determine when the change occurred.
JMP has a function to detect change points in multivariate numerical data, showing when the change occurred using statistical plots. This works by determining the point at which a change can be considered to have occurred based on the changes in the mean and variability when the time series data is divided into two parts at a certain point in time.
This presentation will explore how to detect points of change using readily available data such as weather data, and examine the background behind these changes. This function is useful not only for detecting shifts in manufacturing process parameters, but also for identifying periods when measurement methods or observation environments changed, providing clues to understanding the events behind the data.
