Features of forecasting the demand for tourism services based on a linear model

In the quantitative method of an economic indicator forecasting, an effective approach is to initially assess the state of the time series. The time series in question may have an uptrend, downtrend or sideways trend, i.e. flat. Next, if we have an uptrend or downtrend, then we build a linear trend model. In this case, using the least squares method, one should calculate the condition number of the residual equations matrix. If the condition number is too large (for example, millions, as in the above-mentioned example), then the residual equations matrix should be simplified, for example, by entering centered coordinates. When calculating the tangent of the slope of the model, it is advisable to center only the abscissas and to use the original ordinates. The solution should be checked.