Using a Normalized Score Multi-Label KNN to Classify Multi-label Herbal Formulae
The popularity of herbal medicines has greatly increased in worldwide countries over recent years. Herbal formula is a form of traditional medicine where herbs are combined to heal patient to heal faster and more efficiency. Herbal formulae can be divided into categories. Some formulae can be classified as more than one category. The categories are usually based on indications of herbs in formulae. To support experts for classifying a formula to one or more therapeutic categories, the normalized score multi-label k-nearest neighbors (NSML k-NN) algorithm, is proposed for multi-label herbal formulae classification. The k-NN classifiers with several term weight schemes are explored. The normalized scores are calculated. The values of k, strategies to assign categories are investigated to adjust the decision for multi-label herbal formulae. The experiment is done using a mixed data set of herbal formulae collected from the Natural List of Essential Medicine and the list of common household remedies for traditional medicine. Moreover, a set of well-known commercial products are used for evaluating the effectiveness of the proposed method. From the results, the NSML k-NN is an efficient method to classify multi-label herbal formulae.
KeywordsMulti-label document text classification text categorization herbal formula k-NN classifier
Unable to display preview. Download preview PDF.
- 1.Lovell-Smith, H.D.: In defence of ayurvedic medicine. The New Zealand Medical Journal 119, 1–3 (2006)Google Scholar
- 3.Roiger, R., Geatz, M.: Data Mining: A Tutorial Based Primer. Addison-Wesley, Boston (2002)Google Scholar
- 6.Duwairi, R., Al-Zubaidi, R.: A hierarchical k-nn classifier for textual data. The International Arab Journal of Information Technology 8, 251–259 (2011)Google Scholar
- 10.Singhal, A., Salton, G., Buckley, C.: Length normalization in degraded text collections. Technical Report TR95-1507 (1995)Google Scholar
- 13.Fujino, A., Isozaki, H., Suzuki, J.: Multi-label text categorization with model combination based on f1-score maximization. In: Proceeding of The 3rd International Joint Conference on Natural Language Processing, pp. 823–828 (2008)Google Scholar
- 14.Hua, L.: Research on multi-classification and multi-label in text categorization. In: Proceeding of International Conference on Intelligent Human-Machine Systems and Cybernetics, pp. 86–89 (2009)Google Scholar
- 16.Younes, Z., Abdallah, F., Denœux, T.: An Evidence-Theoretic k-Nearest Neighbor Rule for Multi-label Classification, pp. 297–308 (2009)Google Scholar