Heterogeneous medical treatment data mining-based thyroid cancer risk prediction method

Abstract

A heterogeneous medical treatment data mining-based thyroid cancer risk prediction method is disclosed and relates to the field of thyroid cancer risk prediction. The method is used for solving problems of technologies of the prior art such as insufficient medical treatment data models, poor capability of a learning model for discovering unknown categories, affected anticipation capability of a disease risk prediction model and the like. The method comprises the following steps: a step of collecting medical data and constructing a heterogeneous medical record information network model; a stepof establishing the learning model based on unknown category discovery; a step of achieving learning of unlabeled medical data; a step of constructing an interval qualitative network as a medical reasoning model based on a qualitative Bayesian medical reasoning model and verifying a prediction conclusion. Via the method, all types of semantic information and multiple relationships included in datarecords of patients in different time dimensions can be accurately described; based on a semi-supervised prediction model capable of discovering unknown categories to enable the learning of large amounts of unlabeled medical data, and a problem of predicting disease risk factors can be solved; at last, the method comprises a step of two-way reasoning, an inferred result has positive and negativepolarities, and cause and effect intensity can be expressed by an interval value.

Description

technical field [0001] The invention relates to a thyroid cancer risk prediction method based on heterogeneous medical data mining. Background technique [0002] With the digital accumulation of electronic medical records and electronic health records, medical big data research has been highly valued by researchers in the medical and computer fields. Medical data itself combines the four basic characteristics of big data, namely, a large number, diversity, and rapidity, and generates value, and it also has the characteristics of variability, accuracy, complexity, and heterogeneity. Medical big data contains a wealth of medical knowledge, some of which have not yet been recognized by the medical community. Using this knowledge can not only assist medical treatment and improve medical quality, but also predict medical phenomena and effectively prevent and control diseases. Traditional medicine is the judgment and decision-making of small data, which completely depends on the ...

AI Technical Summary

Problems solved by technology

[0017] The present invention provides a thyroid cancer risk prediction method based on heterogeneous medical data mining to solve the problems in the prior art that the medical data model is insufficient, the learning model has poor ability to discover unknown categories, and affects the prediction ability of the disease risk prediction model.

Images