Time-Series Forecasting

Chapter 13 Time-series forecasting

The core work of time-series forecasting was that only rely on the past observations of that quantity and makeing predict occurrences. The result of time-series forecasting is not considered as the copy of the past but seemed as the past similar to the future. Forecasts are important tools in decision making and can be used in many areas such as marketing, financial, operations, and human resources. There’s always a time horizon in the forecast work which used to indicate how far could be predict. In this chapter, the author tell us a short-term forecasts and use many time-series methods as simples. Patterns are also important in the forecast work, seasonality and trend are two main classses of such patterns. At the last of this chapter, other advanced techniques also been introducted. There’s also some considerations when preparing and using a forecast are take into discussion at the end of the chapter.

Forecasts are useful especially when planning and making a decision in financial business situations. People usually do forecast work in order to predict future data rely on the past data. In this chapter, the author first use a simply case to explain the one-period forecats. And then, in order to evluate each forecasting approaches, the author make a comparison between these results and find out some error or relative error to analysis. Next, discussion about the sasonality and trend patterns in the data is the main content of exploiting multiperiod patterns part. However, finally when implementing these methods , we should considerate more and must have several task to help with using a forecast. In the last part of this chapter “choose a technique and monitor it, decide on aggregation, determine initial model parameters, use the forecasts in decision making, include qualitative judgments, and monitor forecast accuracy” these situation are being explained in detail.