Classification of forecasting methods and models

I have been doing time series forecasting for over 5 years now. Last year, I defended my thesis on the topic “ Time Series Forecasting Model for the Sampling of Maximum Similarity ”, but the issues after defense remained decent. Here is one of them - the general classification of methods and forecasting models .

Usually in the works of both domestic and English-speaking authors do not ask the question of classifying methods and models of forecasting, but simply list them. But it seems to me that today this area has grown and expanded so much that even the most general, but classification is necessary. Below is my own version of the general classification.

What is the difference between a method and a forecasting model?

The forecasting method is a sequence of actions that must be performed to obtain a forecasting model. By analogy with cooking, a method is a sequence of actions, according to which a dish is prepared - that is, a forecast will be made.

The forecasting model is a functional representation that adequately describes the process under study and is the basis for obtaining its future values. In the same culinary analogy, the model has a list of ingredients and their ratio, which is necessary for our dish - the forecast.

The combination of the method and model form a complete recipe!

Currently, it is customary to use English abbreviations of the names of both models and methods. For example, there is a famous prediction model autoregression of an integrated moving average taking into account an external factor (auto regression integrated moving average extended, ARIMAX). This model and its corresponding method are usually called ARIMAX, and sometimes the Box-Jenkins model (method) by the name of the authors.

We first classify the methods.

If you look closely, it quickly turns out that the concept of “ forecasting method ” is much broader than the concept of “ forecasting model ”. In this regard, at the first stage of classification, methods are usually divided into two groups: intuitive and formalized [1].

If we recall our culinary analogy, then there it is possible to divide all the recipes into formalized, that is, recorded by the number of ingredients and the method of preparation, and intuitive, that is, never written down and obtained from the culinary experience. When do we not use the recipe? When the dish is very simple: fry potatoes or cook dumplings - here the recipe is not needed. When else do we use the recipe? When we want to invent something new!

Intuitive forecasting methods deal with judgments and expert evaluations. Today, they are often used in marketing, economics, politics, as a system whose behavior needs to be predicted is either very complex and not amenable to mathematical description, or very simple and does not need such a description. Details about such methods can be found in [2].

Formalized methods are the forecasting methods described in the literature, as a result of which they build prediction models, that is, they determine a mathematical relationship that allows one to calculate the future value of the process, that is, to make a prediction.

This is where the general classification of forecasting methods, in my opinion, can be completed.

Next, we make a general classification of models.

Here it is necessary to proceed to the classification of forecasting models. At the first stage, the models should be divided into two groups: domain models and time series models.

Domain models are mathematical prediction models that use domain laws to build them. For example, the model on which the weather forecast is made contains the equations of fluid dynamics and thermodynamics. The forecast of population development is done on a model built on a differential equation. The prediction of the blood sugar level of a person with diabetes is made on the basis of a system of differential equations. In a word, in such models dependencies are used that are characteristic of a specific subject area. This kind of model is peculiar to an individual approach to development.

Time series models are mathematical forecasting models that seek to find the dependence of the future value on the past within the process itself and on this dependence calculate the forecast. These models are universal for different subject areas, that is, their general appearance does not change depending on the nature of the time series. We can use neural networks to predict air temperature, and then use a similar model on neural networks to predict stock market indices. These are generalized models, like boiling water, in which if you throw a product, it will be cooked, regardless of its nature.

We classify time series models.

It seems to me that it is not possible to make a general classification of domain models: how many regions, so many models! However, time series models are easy to divide easily [3]. Time series models can be divided into two groups: statistical and structural.

In statistical models, the dependence of the future value on the past is given in the form of some equation. These include:

1. regression models (linear regression, non-linear regression);
2. autoregressive models (ARIMAX, GARCH, ARDLM);
3. exponential smoothing model;
4. maximum sampling model;
5. etc.

In structural models, the dependence of the future value on the past is given in the form of a certain structure and rules for transition along it. These include:

1. neural network models;
2. Markov chain models;
3. models based on classification and regression trees;
4. etc.

For both groups, I indicated the main, that is, the most common and described in detail forecasting models. However, to date, there are already a huge number of time series forecasting models and, for example, SVM (support vector machine) models, GA (genetic algorithm) models and many others have been used to build forecasts.

General classification

Thus, we obtained the following classification of models and forecasting methods .

References

1. Tikhonov E.E. Forecasting in market conditions. Nevinnomyssk, 2006. 221 p.
2. Armstrong JS Forecasting for Marketing // Quantitative Methods in Marketing. London: International Thompson Business Press, 1999. P. 92 - 119.
3. Jingfei Yang M. Sc. Power System Short-term Load Forecasting: Thesis for Ph.d degree. Germany, Darmstadt, Elektrotechnik und Informationstechnik der Technischen Universitat, 2006. 139 p.

UPD. 11/15/2016.
Gentlemen, it came to insanity! Recently, I was sent for review an article for the VAK edition with reference to this entry. I draw your attention to the fact that neither in diplomas, nor in articles, much less in dissertations can one refer to a blog ! If you want a link, use this: Chuchueva I.A. MODEL FORECASTING TEMPORARY ROWS FOR THE SAMPLE OF MAXIMAL SIMILARITY, thesis ... Cand. those. Sciences / Moscow State Technical University. N.E. Bauman. Moscow, 2012.