(Statistics) An iterative process for developing a model beginning with some information about the form and structure of the problem and with relevant data. The model building process typically follows a sequence of inter-related steps to include: (1) Problem Identification and Data Selection: Data is selected, compilation, screened, and analyzed, and the various series tested based on hypotheses of probable causation; (2) Model Identification (or Specification): Selection of a general model structure is made based on the nature of the data and the types of outputs desired. Some of these include, for example, a simple single mathematical equation, or multiple (sequential) equations, statistically-based univariate (deterministic) autoregressive functions, multivariate analysis, simple ordinary least squares (OLS) regression, multiple regression, simultaneous equation, etc.; (3) Estimation (Model Fitting): Based on the selection of a model structure, the data is used to best describe the behavior of the variable under observation, e.g., stream flows, reservoir levels, runoff, economic output, employment, consumer spending, etc.; (4) Model Testing (and Refinement, as Necessary): The model's structure and variables chosen are then validated by applying the data and observing forecast errors with respect to know (sample) values; (5) Forecasting: Based upon the ability of the model to accurately "fit" or predict historical values, the model is used to forecast beyond the last data point as prescribed by scenarios under analysis. Also see Econometrics, Regression Analysis, Stochastic Process, and Deterministic Process.

Environmental engineering English vocabulary.      Английский словарь экологического инжиниринга.