Domain Specific Classification of Malay Based Complaints using the Complaint Concept Ontologies

Shaiful Bakhtiar bin Rodzman, Siti Suhaima binti Suhaili, Normaly Kamal Ismail, Nurazzah Abd Rahman, Syed Ahmad Aljunid, Aslida binti Omar · 2019

The complaint from users is an effective method to identify the quality of services and facilities provided by an organization. The efficiency to respond to users' complaint also depends on an effective workflow. By having an effective method and workflow, the action taken by the management to improve the quality of services and facilities can be done immediately and effectively. One of the ways is by classifying the complaints that will isolate related complaints. This paper presents the implementation of the classification system that combines the application of Complaint Concept Ontologies in Malay language as classifier rules with the BM25 model of Information Retrieval system. Experiments showed the semantic based elements such as Malay ontology may bring the improvement of the classification of the Malay Complaint. The result yielded showed that the proposed classifier produced better result in four category compared to BM25 original score that only produced better result in one category. OBMCS also outperformed the LDA model in all eight categories on the Recall, Precision and F-measure metrics. The finding proven the proposed system is very useful, especially to the Malay complaint in regards of classification for documents in the domain area.

Read the paper · More papers on PaperTik