An Applied Data Mining Technique to Develop on Information Storage and Retrieval for Intelligence of Locality for the Southern Part of Thailand in Suratthani Province
This study was conducted for the research and collection of the articles about local wisdom in Suratthani Province, Southern Thailand which were published on Google website. The searched articles were collected, analyzed, and categorized by K-Means Cluster Technique. The analysis results indicated that the data can be categorized into 15 groups due to the statistical significant value of 0.000 and the sum square error value of 0.81 which was near zero. The data of cluster were concentrated which was reliable. The clusters were congruent with the article contents and the frequency of keywords. All articles were analyzed based on the association rule and categorized by the formal context analysis resulting in 15 objects with 245 attributes. The analysis results generated into 319 association rules relating to the keywords of articles. There were keyword overlaps in 45 articles with the edge count of 87 articles. The highest keyword overlaps were in 6 articles. The results of formal contcxt analysis of 15 groups could be used for category analysis. The categories and the association rules could be used for design, analysis, and development of information storage and retrieval system. This consists of the category system, the article creating system, the metadata related to article system, and the retrieval system. The efficacy of the system and the precision ratio of the prototype was 98% or at the highest level, while the recall ratio of the prototype was 81% or at high level.