Julia for Data Science by Zacharias Voulgaris PhD

Author:Zacharias Voulgaris PhD [Zacharias Voulgaris PhD]
Language: eng
Format: epub
Tags: Big Data, Digital Media, Math & Science
Publisher: Technics Publications
Published: 2016-08-31T21:00:00+00:00

g

0

11747

585

12332

h

34

4778

1876

6688

Total

34

16525

2461

19020

Table 7.2 A contingency table showing counts of “Low”, “Medium,” and “High” values of the fLength feature, against the class variable.

Based on the above data (which is stored in, say, the two-dimensional array C, comprising of the values of the first 2 rows and the first 3 columns of the table), we can apply the chi-square test as follows:

In[32]: ChisqTest(C)

Julia will spit out a lot of interesting but ultimately useless information regarding the test. What you need to focus on is the summary that’s buried somewhere in the test report:

Out[32]: Test summary:

outcome with 95% confidence: reject h_0

two-sided p-value: 0.0 (extremely significant)

Clearly there is no way that these counts could have been generated by chance alone, so we can safely reject the null hypothesis (which Julia refers to as h_0) and conclude that there is a significant relationship between fLength and the class variable. Again, this means that we may want to include it (or the original feature it was created from) in our model.

The result of the chi-square test will not always be so crystal clear. In these cases, you may want to state beforehand what your significance threshold (alpha) will be and then compare the p-value of the test to it. If it’s less than the threshold, you can go ahead and reject the null hypothesis.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.