I recommend 4CF as a dependable partner in Strategic Foresight and Futures Literacy for policy-making and Strategy.
Project results as well as the process whereby they were obtained have me convinced as to the professionalism and innovativeness of 4CF on the Polish market.
4CF provided foresight expertise at a highest level of professional skill.
4CF experts successfully implemented a Real-Time Delphi tool - HalnyX - for the needs of an expert study.
4CF offers the highest quality of Business Foresight services
Applications: Wherever quality data is available, statistical modeling can be used. This method, however, should be treated as a benchmark for other scenarios or visions of the future. Ending futures research on modeling can be deceptive.
Method: Today's statistical modeling and forecasting methods have been developing since the 1950s. It started with demand forecasting, with the use of tools such as moving averages and exponential smoothing. It later moved to more advanced methods - based on the assumption that demand has its structure and regularity (e.g. consumer behavior patterns) - such as Winter's averaging method, Fourier series, or the Box-Jenkins method (ARIMA) (Lapide, 1997). Finally, a time came for advanced analyses of the impact of a company’s specific actions (marketing, price policy) in the form of econometric models . The first of these were simple, but they gradually became more advanced and began to use statistical and forecasting software based on multiple regression, neural networks, or genetic algorithms (Lapide, 1997).
Currently, statistical modeling methods can be divided into two main branches:
- Based on time series. These extrapolate historical data with the use of the method which is the best statistical match. We assume that the regularity present in historical data will also occur in the future (moving average, exponential smoothing, trend, decomposition, Box-Jenkins method (ARIMA)),
- Based on cause-effect modeling. These include econometric models in which it is crucial to define the cause (called the independent / explanatory variable) and the effect (called the dependent / explained variable). The value of the dependent variable - as the name implies - depends on the values of the independent variable(s). This branch encompasses methods ranging from simple regression to more complex econometric models and methods based on neural networks (Jain, 2006a). Cause-effect models also include models based on system dynamics.
In practice, we can also talk about a hybrid method - the "decision and forecast" tree - in which forecasts of variables are the basis for further predictions and so forth, until the final forecast necessary for the decision-making process is obtained. Each variable can be obtained by means of a different method. It might also be necessary, at times, to use several methods in order to predict one variable. The final forecast can sometimes be obtained through, for example, averaging values obtained by various means: from linear trend estimation, autoregressive and econometric models.
4CF has carried out a number of successful statistical projects that have made it possible to optimize public policies and strategies. We are one of very few companies worldwide to base our strategic foresight on a strong quantitative foundation, integrating mathematical methods, social research and psychology and supporting all of the above with analytical engines and other IT innovations.
Statistical modeling can also be utilized to decompose sales affecting factors, you can read about it here.