skip to main |
skip to sidebar
This model is useful to study the relationships between two variables (explanatory and explained), for example: amount depending on the price, sales in terms of advertising, production as a function of the number of hours worked; demand-price and demand-rent, ...
Once you have chosen the functional relationship with the best fit (linear, potential, exponential, logarithmic or parabolic) , this template can be used for simulations and future estimations.
This Excel template predicts trends, creating a mathematical function of time by simple regression techniques. In other words, trying to explain a dependent variable (also called endogenous) according to the dependent variable time (exogenous or explanatory). After entering data we can see the graph and trends calculated, choosing the model with a greater degree of fit (coefficient of determination higher). The selected function can be used to make predictions of the variable under study (y), assigning values in function to the time variable (x). It can also be seen in the graph, where there is the trend by adding more data of time (years, months, ...) in the DATE column. The estimation is useful for time series forecasting (economic or otherwise) when all you know is the data series over time.
Exponential smoothing is obtained by a moving average with weights declining in geometric progression. Brown's double exponential smoothing, includes a trend component which is useful for forecasting series with or without trend but no seasonality. To achieve optimal prediction we need an ALFA coefficient that minimizes squared error, which is achieved by trying different values between zero and one in the following Excel model.
This technique is used to forecast economic series such as: sales, revenues, costs, expenses, production, wages, economic indicators...
