A new HGA-FLVQ model for Mycobacterium Tuberculosis detection

This research aims at develop an MTB detection model from the FLVQ neural network to HGA-FLVQ model. In this research, the FLVQ method was developed through strengthening its initiation, in which the first cluster centers used as FLVQ input were optimized first by HGA. The results show that sensitivity and specificity of the HGA-FLVQ model reach 96.30 and 95.65%, whereas the sensitivity of an FLVQ method is 70.83%, and the sensitivity of an LVQ method is 87.50%. The specificity of an FLVQ method and the specificity of an LVQ method are 84.62%. Based on these results, we can say that the HGA-FLVQ model is better than FLVQ and LVQ methods. It also means that relative amplitude can be used by the HGA-FLVQ model as a feature to detect the presence of MTB in the sputum of TB-suspected patients. Thus, the HGA-FLVQ model can be used to strengthen TB laboratory examination at Public Health Centers in Indonesia.