The following information was submitted:
Transactions: WSEAS TRANSACTIONS ON BIOLOGY AND BIOMEDICINE
Transactions ID Number: 53-713
Full Name: Doungporn Maiprasert
Position: Student
Age: ON
Sex: Female
Address: Rangsit university
Country: THAILAND
Tel: -
Tel prefix: -
Fax: -
E-mail address: dmiprasert@yahoo.com
Other E-mails: dniyomua@yahoo.com,kr_stat@yahoo.com
Title of the Paper: prediction of the stages of breast cancer using principal component analysis â€"a comparison of artificial neural network with multinomial logistic regression
Authors as they appear in the Paper: Doungporn Maiprasert, Krieng Kitbumrungrat
Email addresses of all the authors: dmiprasert@yahoo.com,kr_stat@yahoo.com
Number of paper pages: 10
Abstract: Breast cancer is a disease that results in a loss of lives of many Thai women. The objective of this research is to develop a method for predicting stages of the cancer before it reaches the final stage of cancer. In this research, the analysis and prediction of the stages of cancer by using Principal Component Analysis - Artificial Neural Network (PCA-ANN) and Multinomial Logistic Regression (MLR) was performed. The two methods are different; the method of PCA-ANN helps reduce the features or symptoms in the prediction but the method of Artificial Neural Network uses backpropagation learning. The PCA-ANN can predict the accuracy of overall by 82.6% and the MLR method can predict the cancer risk. The results of the analysis revealed that MLR can predict the overall accuracy by 74.1%. However, both methods have different advantages. The MLR can predict the cancer stages, especially the benign stage and stage 4. The PCA-ANN method can predict all stages. In the eva!
luation of the prediction, we used two values: sensitivity and specificity. The PCA-ANN method gives more sensitivity value than the MLR method but the MLR method gives more specificity value than the PCA-ANN method.
Keywords: Principal component analysis, Artificial neural network, Multinomial logistic regression, Backpropagation, Sensitivity, Specificity
EXTENSION of the file: .pdf
Special (Invited) Session: Mathematical and Computational Methods in Biology and Biomedicine
Organizer of the Session: Radoslav S. Bozov
How Did you learn about congress: prediction
IP ADDRESS: 202.28.120.21
