Practice 1

Practice 1 is an example of linear regression. This means we could use some method to imitate the trend of data, then could use the trend to predict the future potential value.

Practice 2

Practice 2 is an example of data optimization. The first thing for this kind of task is to define the target. In this case, the target is to get the minimum value of the optimization cell. The future task may be to maximize the target or approaches a specified value.

The time series data normally consist with multi-components, the most two important components are trend and seasonality, the trend could be using leaner regression to get

The final result

The time series data normally consist of multi-components, the most two important components are trend and seasonality, the trend could be modeled using linear regression to get, and the seasonality could use trigonometric functions to model. No matter trend or seasonality they are all stationary relationships. The residuals will get after removing the trend and seasonality, which are the main research objects. If could get a stationary relationship from the residuals, it would be easier to predict.