Statistics Panel

To show statistics panel:

• Click the Statistics button in the Working Area group on the View ribbon tab.

• Select the Statistics checkbox in the context menu of the data view area.

The panel appears containing characteristics of the object selected in the data view area.

To hide statistics panel:

• Release the Statistics button.

• Deselect the Statistics checkbox in the context menu of the data view area.

The statistics panel is hidden.

The tab is available only for the calculated series calculated using modeling methods (smoothing, forecasting, or regression).

The tab shows identified equation with estimated coefficients and a coefficient table containing the following data:

• Factor. Name of the factor, by which the coefficient was estimated. It is displayed only for factor modeling methods.

• Coefficient. Name of the factor of the estimated coefficient.

NOTE. If on the Parameters side panel tab, options for counting order of non-seasonal or seasonal autoregression and/or non-seasonal or seasonal moving average are enabled for the ARIMA method, AR/MA and/or SAR/SMA coefficients are displayed for this method.

• Initial Approximation. Calculated initial approximation.

NOTE. It is displayed only for the Non-Linear Regression calculation method.

• Value. Estimated coefficient value.

• Standard Error. An approximate value of deviation (caused by the randomness of the sample) of the coefficient estimation from the actual value. The greater is the value of the standard error, the less reliable is coefficient estimation.

• t-statistics. It is calculated as a ratio of coefficient estimation to its standard error. It is used to test the null hypothesis that the coefficient is equal to zero, and is therefore insignificant in the analyzed regression model. The hypothesis assumes that the residuals have normal distribution. It is easier to interpret value of the t-statistics using probability value given next to it.

• Probability. Probability value for t-statistics. The null hypothesis on the estimated coefficient equality to zero is rejected at the selected importance level (usually 0.1, 0.05, 0.01), if the probability value is less than the importance level.

Features of data display

• For the trend forecasting methods the tab also displays the name of the dependence type. For example:

• For the Exponential Smoothing method the tab displays only identified equation with estimated coefficient and table of coefficients containing only their values.

The tab displays summary statistics of the selected series. Various set of characteristics can be displayed for different series. If the selected range of data belongs only to the calculated series and is obtained using modeling methods (smoothing, forecasting, or regression), the set of characteristics is similar to the summary statistics displayed in the modeling container:

Detailed description of summary statistics is given in the Library of Methods and Models section.

The tab is available only for the calculated series calculated using the Linear Regression modeling method.

The table shows values of correlation coefficients for explanatory series. The cells are highlighted with appropriate colors:

• Yellow. Medium correlation. Absolute value of the correlation coefficient is in the range from 0.5 to 0.75.

• Green. Strong correlation. Absolute value of the correlation coefficient is in the range from 0.75 and higher.

The tab is available only for calculated series calculated by the Baxter-King Filter method.

The tab contains weight matrix used by the Baxter-King Filter method. The columns contain calculated coefficient values.

Matrix dimensions: 1×(q+1), where q is the value of Lead/Lag parameter specified on the Parameters tab.