Econometrics in tasks and exercises. Demeshev B.B., Borzykh D.A.

The proposed reader of the manual contains the tasks of the most different levels of complexity - from the typical ones, which are given with a detailed solution, to tasks of increased complexity. In addition to the usual tasks that can be solved “manually”, a large number of tasks are selected that require the use of any statistical packages. The tasks that develop programming skills in these packages are given.

The problem covers all the main sections on the introductory level:

• Assessment of linear regression models according to the method of the smallest squares (MNK); /li>
• Gaussa-Markov Theorem and the properties of MNC-fixes;
• Building trusting intervals for the parameters of these models;
• Testing hypotheses;
• Features of evaluation linear regression models under the conditions of violation of the Gauss-Markov Theorem (case of multi-liners, heteroscedicism and autocorrelation);
• tests for the correct functional shape of the model (Ramsey test and boxing test).

In addition, the manual contains tasks on topics that belong to the intermediate level:

• Assessment of economies by maximum credibility (MMP);
• Testing of hypotheses using three classic tests associated with the method of maximum credibility: the Test Relations of credibility, the test of Lagrange multipliers and the Walda test;
• Model of Binar choice (Logit and Probit models);
• Model with random regressors ;
• Elements of the theory of time rows (ARMA processes, the Dick Fuller test for the stationary row, Garch-models).

There are also two sections of an advanced level in the task : the method of support vectors and Random Forest, and two auxiliary sections that contain the necessary information on linear algebra and probability theory.