MTH-416, REGRESSION ANALYSIS Lecture Notes - Lecture 25: Indian Institute Of Technology Kanpur, Multicollinearity, Time Series
6 Apr 2023
Department
Professor
Document Summary
The number of condition indices that are large, say more than 1000, indicate the number of near-linear dependencies in. A limitation of cn and jc is that they are unbounded measures as 0 cn . X x jc : measure based on characteristic roots and proportion of variances: X x obviously, v is an orthogonal matrix. Then is a k k matrix and v is a k k . 2 k k dependency in the data, then j is close to zero and the nature of linear dependency is described by the elements of the associated eigenvector jv . v k v are the elements in ik. C jc is greater than the danger level 1000. "i s for which (b) this gives the number of linear dependencies. (c) don"t consider those. "jc s which are below the danger level. (iii) " s with condition index above the danger level, choose one such eigenvalue, say (iv)
