Thai-English spam SMS filtering

SMS spam filtering for Thai-English language has not previously been studied and implemented. Two methods of spam SMS message filtering objected to filter spam SMS messages written in Thai and English have been studied and implemented. The first method simply uses current spam English message filtering and then upgrades for Thai language support. The second one applies text normalization, word segmentation process, and analyzing/correcting the semantic of Thai words. However, both methods are applied by 2 different decision-making algorithms: Support Vector Machine (SVM) and Naive Bayesian (NB) algorithms. Finally, the results from trial applying in the real system are shown. The results show that the second filtering method has higher accuracy. Moreover, the SVM based filtering consumes more processing time than the NB based filtering about 2.5 times for both proposed methods.