TY - JOUR AU - GANGULY, ANINDYA AU - BISWAS, TAPAN KRISHNA AU - BANDOPADHYAY, RAJIB AU - DUTTA, AJOY KUMAR PY - 2018/07/31 Y2 - 2024/03/29 TI - APPLICATION OF PRINCIPAL COMPONENT ANALYSIS TO CLASSIFY NORMAL BRAIN TISSUE AND BRAIN LESIONS LIKE LOW AND HIGH GRADE GLIOMA, METASTASES AND MULTIPLE SCLEROSIS JF - IJRDO - Journal of Electrical And Electronics Engineering JA - eee VL - 4 IS - 7 SE - Articles DO - 10.53555/eee.v4i7.2168 UR - https://www.ijrdo.org/index.php/eee/article/view/2168 SP - 20-34 AB - Principal Component Analysis ( PCA) an extremely useful method of Statistical techniques is applied when working with a lot of parameters or independent numerical variables to predict the different pathological lesions in the brain like Multiple Sclerosis (MS), Glioma , Glioblastoma of different grades and Metastasis. Statistical techniques such as factor analysis or Principal Component Analysis(PCA) help to overcome such difficulties.In different brain diseases structural alterations in the normal tissue may be noticed in MR images. It is not so simple to detect the brain lesions correctly even from the MR spectroscopic graph. Enormous data collected from various patients such as – Refractive Index, T2 relaxation values, Apparent Diffusion Coefficient (ADC), Creatine (CR), Choline (CHO), NAA (N-Acetyl Aspartate), ratio of CR/NAA, LIP/LAC (Lipid/lactate), MI ( Myoinositol), CHO/CR and T2 value in the periphery of lesion may be confusing. The relationship between each variable may not be clear and that there is a chance of over fitting the data. By reducing the dimension of the feature space by “feature elimination” and “feature extraction”, there may be less chance of over fitting the data. PCA helps identifying the disease condition in doubtful cases by generating a map depicting and classifying the diseases. ER -