Breast cancer risk evaluation analysis system based on deep convolution neural networks

Abstract

The invention relates to a breast cancer risk evaluation analysis system based on deep convolution neural networks. The breast cancer risk evaluation analysis system comprises a medical document pretreatment module, a word vector training module, a distributed semantic feature medical information extraction module, a long-term semantic associated feature extraction module and a breast cancer risk evaluation analysis module. The medical document pretreatment module is used for pre-treating medical text big data to generate a word table for word vector training; the word vector training module generates a primary word vector by training a deep convolution neural network; the distributed semantic feature medical information extraction module maps learned distributed feature representation to sample marker space through a fully connected layer and generates distributed semantic features of medical fields; the long-term semantic associated feature extraction module extracts long-term semantic associated features of medical clinical documents by the aid of the distributed semantic feature representation; the breast cancer risk evaluation analysis module trains a deep convolution neural network for breast cancer risk evaluation by the aid of the long-term semantic associated features and evaluates breast cancer risks. The system improves automation and intelligence level of screening of breast cancers.

Description

technical field [0001] The invention relates to the technical field of medical equipment, in particular to a breast cancer risk assessment and analysis system based on a deep convolutional neural network. Background technique [0002] In recent years, the incidence of breast cancer in our country has been increasing year by year, especially in some big cities, such as Shanghai, Beijing and other places, breast cancer has jumped to the first place in the incidence of malignant tumors in women. Massive structured and semi-structured data and intricate unstructured data challenge the medical industry, making it difficult to rationally allocate resources and putting tremendous pressure on the development of the entire medical industry. In the case of breast cancer, patients' electronic medical record information is scattered in narrative medical text, but most computer applications can only understand structured data. A common practice is to use machine learning methods to docu...

AI Technical Summary

Problems solved by technology

These evaluation indicators are called characteristics, and are usually oriented by specific disease types, which can easily lead to over-engineering and do not have wide applicability

Images