Statistical Sciences 2035 Study Guide - Quiz Guide: Empty Spaces, Principal Component Analysis, Scree
7 Oct 2018
School
Department
Professor
Document Summary
We limit the number of factors to those that are important by choosing factors with the largest eigenvalues. The simplest method is to choose those dimensions with eigenvalues that are 1 or greater because they explain more variation than did an original item. The selection of factors for interpretation with eigenvalues greater than 1 is known as kaiser s rule. An alternative method called the scree test where successive eigenvalues are plotted on a graph, and the spot where the plot abruptly levels out or goes below one is the cut- off point. In general, there is a four-stage sequence to follow in conducting a factor analysis: descriptives and a correlation matrix are generated for all the variables. The first table of descriptives shows the means, standard deviations and ns. In the second table, correlation matrix, first indicators whether a pca is worth to conduct can be observed. Usually, there have to be some correlation coefficients higher than.
