Time series is a term used to refer a sequence of values that are observed over a period of time in a chronological order. Time series is a set of data that is not counted as a branch of science. In time series, even if a data set tells the previous value of series it is very difficult to predict the certainity level of the next value of any variable. We say that the series is random or non-deterministic and these certainly is of concern to the body of doctrine known as the time series analysis and we're going to dedicate this brief introduction. The objectives of time series analysis are diverse and can highlight the prediction, process control, process simulation, and the generation of new theories of physics or biology.
An important concept found in this area is that of stationary processes. For instance, if we look at temperature for a given month over the years in a given geographical area, and climate change is happening, although there are fluctuations, there will be a growing trend. Informally, we say that a series is stationary trend when it is in statistical equilibrium in the sense that their properties do not vary over time, and therefore trends cannot exist. A process is difference-stationary if its properties vary with time, as the weather.
For the given data set we would be applying the ARIMA modeling of Time series analysis, this model represents a family of models characterized by three parameters (p, d, and q) that describe the basic properties of a specific time-series model. The value of the first parameter p denotes the order of the autoregressive component of the model. If an observation can be influenced only by the immediately preceding observation, then the model is of order one. If an observation can be influenced by the two immediately preceding observations, then the model is of order two, and so on. The value of the second parameter d refers to the order of differencing that is necessary to stabilize a difference stationary time series. A process is described as difference stationary when the values do not vary about a fixed mean level; rather, the series might first fluctuates about one level for some observations and fluctuate about a different level. The value of the third parameter q denotes the order of the moving average component of the model. Again, the order describes how many preceding observations must be taken into account. Higher order models (p or q > 3) are rare in the behavioral sciences.
Discussion
The time-dependent data, as a rule, are unpredictable and cannot be modeled or predicted. Therefore, the results obtained with the use of difference stationary trend, may be artificial - they can point to a relationship between two variables, where one basically does not exist. To receive the consistent, reliable results, time-dependent data must be converted to fixed data. In contrast to the difference-stationary process, which is variable and the ...