Skip to main content

Table 2 A list of rules extracted from 324 tags using the Apriori algorithm.

From: Machine learning approaches to supporting the identification of photoreceptor-enriched genes based on expression data

Enriched in PR? <= criterion 4 (234:72.2%, 0.829)
Enriched in PR? <= criterion 1 (218:67.3%, 0.899)
Enriched in PR? <= criterion 3 (177:54.6%, 0.842)
Enriched in PR? <= criterion 2 (175:54.0%, 0.937)
Enriched in PR? <= criterion 4 & criterion 1 (165:50.9%, 0.909)
Enriched in PR? <= criterion 4 & criterion 3 (129:39.8%, 0.884)
Enriched in PR? <= criterion 4 & criterion 2 (117:36.1%, 0.949)
Enriched in PR? <= criterion 1 & criterion 3 (126:38.9%, 0.929)
Enriched in PR? <= criterion 1 & criterion 2 (143:44.1%, 0.944)
Enriched in PR? <= criterion 3 & criterion 2 (105:32.4%, 0.943)
Enriched in PR? <= criterion 4 & criterion 1 & criterion 3 (92:28.4%, 0.946)
Enriched in PR? <= criterion 4 & criterion 1 & criterion 2 (100:30.9%, 0.96)
Enriched in PR? <= criterion 4 & criterion 3 & criterion 2 (70:21.6%, 0.971)
Enriched in PR? <= criterion 1 & criterion 3 & criterion 2 (90:27.8%, 0.944)
Enriched in PR?<= criterion 4 & criterion 1 & criterion 3 & criterion 2 (59:18.2%, 0.966)
  1. Each rule is shown in the following format: Enriched in PR? <= criterion 1 & criterion 2 & ... &criterion n where the rule is interpreted as "for tags that meet criterion 1 through criterion n, they are likely to be PR-enriched genes." The numbers shown at the end of each rule indicate the number of tags to which the rule applies (Support) and the proportion of those tags for which the rule is true (Confidence). Support is reported both as number of tags and percentage of total tags, separated by a colon.