Premium
[P4–512]: DIAGNOSIS OF ALZHEIMER's DISEASE USING A MACHINE LEARNING TECHNIQUE
Author(s) -
Jha Debesh,
Kwon GooRak
Publication year - 2017
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2017.07.674
Subject(s) - artificial intelligence , pattern recognition (psychology) , principal component analysis , computer science , feature extraction , logistic regression , wavelet , sensitivity (control systems) , wavelet transform , discrete wavelet transform , curse of dimensionality , machine learning , engineering , electronic engineering
Figure 1. ReHo map of the study participants. Background: This research article proposes a smart machine learning technique to discriminate normal and Alzheimer’s brain image. Methods:The method is based on following computational technique; the discrete wavelet transform is utilized for feature extraction, the principal component analysis for minimizing the dimensionality of the wavelet coefficients, and logistic regression to classify the reduced features into normal or Alzheimer’s disease. Results:The experiments were carried out on 90 images consisting 5 normal and 85 pathological from Harvard medical school dataset. The proposed system yielded an excellent classification accuracy of 97.80%, sensitivity 98.82%, and specificity of 80.00% with 5Fold cross-validation. Conclusions: In this study, we proposed a new technique to predict Alzheimer’s disease. The proposed technique outperforms 4state-of-the-art-algorithm in terms of accuracy, and sensitivity. Furthermore, our method signifies its effectiveness when compared with the other machine learning approaches.