NotesFAQContact Us
Search Tips
Peer reviewed Peer reviewed
PDF on ERIC Download full text
ERIC Number: EJ1111559
Record Type: Journal
Publication Date: 2007-May
Pages: 23
Abstractor: As Provided
ISSN: EISSN-2330-8516
An Entropy-Based Measure of Dependence between Two Groups of Random Variables. Research Report. ETS RR-07-20
Kong, Nan
ETS Research Report Series, May 2007
In multivariate statistics, the linear relationship among random variables has been fully explored in the past. This paper looks into the dependence of one group of random variables on another group of random variables using (conditional) entropy. A new measure, called the K-dependence coefficient or dependence coefficient, is defined using (conditional) entropy. This paper shows that the K-dependence coefficient is a measure of the degree of dependence of one group of random variables on another group of random variables. The dependence measured by the K-dependence coefficient includes both linear and nonlinear dependencies between two groups of random variables. Therefore, the K-dependence coefficient measures the total degree of dependence and not just the linear component of the dependence between the two groups of random variables. Furthermore, the concept of the K-dependence coefficient is extended by defining the partial K-dependence coefficient and the semipartial K-dependence coefficient. Properties of partial and semipartial K-dependence coefficients are also explored.
Educational Testing Service. Rosedale Road, MS19-R Princeton, NJ 08541. Tel: 609-921-9000; Fax: 609-734-5410; e-mail:; Web site:
Publication Type: Journal Articles; Reports - Research
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A
Grant or Contract Numbers: N/A