Shrink: A Breast Cancer Risk Assessment Model Based on Medical Social Network

Abstract

Breast cancer risk assessment model can assess whether a people is at a high risk of developing breast cancer disease or not and confirm a breast cancer high-risk group. Because the etiology of breast cancer disease is different in different country and region, the existing risk assessment model is only adaptive to certain countries and regions. And the parameters of these models are fixed, so these models have poor generality. Aiming at these problems, the paper puts forward a new breast cancer risk assessment model named as Shrink. Using the idea of social network, Shrink constructs a medical social network to show the similarity among people, and uses group division algorithm to divide the network into breast cancer high-risk group and low-risk group. The parameters of this model can be set according to the needs of the breast census, and these parameters can be directly acquired through questionnaire, therefore Shrink has good generality. Moreover, under the uncertain classification standard, Shrink adopts a new classification method to discover breast cancer high-risk group. Based on the real data from questionnaires, we make experiments in Matlab, and obtain the evaluation index of the model. The experiment proves that the model itself has good evaluation result and is better than classic Gail model.