Decision Model And Method Of Intelligent Auxiliary Diagnosis And Treatment Based On Hybrid Attribute Information Rough Set

Medical security and services are related to people’s life,health and safety.Efficient and convenient medical diagnosis mode and medical service level are an important embodiment of national economic development.The development and maturity of Internet technology has promoted the rapid development of informatization in the medical industry.At present,China’s existing medical services still face problems such as ineffective development of data resources,unreasonable allocation of medical resources,low level of intelligence and unscientific diagnosis and treatment decisions.Therefore,using emerging technologies to develop and utilize existing data resources,improve the efficiency of clinical diagnosis and treatment,and promote scientific and intelligent diagnosis has become the focus of social attention and an important research and development direction,and has great social and practical significance.Based on rough set theory,this paper focuses on three scenarios:medical examination item recommendation before disease diagnosis,intelligent auxiliary disease risk prediction in diagnosis and treatment alternative selection after diagnosis.Considering the problems existing in these three diagnosis scenarios and the limitations of using existing methods,this paper makes innovative research on existing methods and models by combining machine learning,recommendation technology,three-way decisionmaking and multi-attribute decision-making,effectively solve the problems faced by the above three scenarios.Before disease diagnosis,there are unreasonable allocation of medical resources,so this paper proposes a medical recommendation method based on examination items.However,in this recommendation scenario,there are inconsistencies in the data types and multi criteria in recommendation item.While most of the existing recommendation studies are based on single score recommendation,and the application of the existing multi criteria recommendation method has certain limitations.To solve the above problems,this paper constructs a multi criteria three-way medical examination item recommendation model by integrating the methods of hybrid attribute information rough set,three-way recommendation and grey relation analysis.Using the distance measure of different types of data in hybrid attribute information rough set,the inconsistency of sub criteria data types is solved.The grey relation analysis method is used to calculate the grey relation degree of the target object as the overall score of the recommendation item,which solves the problem of how to calculate the overall score of the item in multi criteria recommendation.Finally,the collaborative recommendation and three-way recommendation are used to predict the score of items that are no overall score.Finally,the method is verified effectively by clinical data and realize the personalized recommendation of examination items.In the process of disease diagnosis,using patients’ medical data to identify and judge the disease risk early is not only convenient for patients to prevent and formulate accurate treatment strategies as soon as possible,but also can assist doctors in scientific diagnosis,effectively improve the efficiency of diagnosis and treatment and improve the accuracy of diagnosis.Considering the problems of inconsistent data types and redundant attribute features in patient medical data,an attribute reduction algorithm based on positive domain invariance based on the hybrid attribute information rough set is proposed to extract attribute features with high correlation and less redundancy.In addition,the three-way decisionmaking considers the decision-making cost,divides the objects with insufficient information into the boundary domain,delays the decision-making,improves the accuracy of classification and reduces the cost of misclassification decision-making.Therefore,a KNN three-way classification method based on hybrid attribute information rough set is proposed in this paper to improve the learning ability of classifier.Finally,the gout data in clinical practice are used to verify the method proposed in this paper,so as to realize the intelligent risk prediction of gout and provide medical knowledge discovery.After the diagnosis of the disease,the selection of treatment alternative is a key and important segment.In the existing mode of treatment alternative selection,the medical service mode dominated by doctors and obeyed by patients is often adopted.This mode only considers the subjective opinions of doctors and ignores many influencing factors such as curative effect,cost and adverse reactions.In addition,the evaluation criteria of treatment alternative often have varied scales,not just a single digital scale.And decision-makers have strong loss aversion and loss avoidance psychology in the selection of treatment alternative.Therefore,this paper proposes a treatment selection model integrating hybrid attribute information rough set and TODIM method,which considers the psychological behavior of decisionmakers and solves the problem of data scale diversification in the evaluation criteria,and realizes the effective selection of treatment alternative.Based on the three scenarios of examination item recommendation before diagnosis,intelligent aided disease risk prediction in diagnosis and treatment alternative selection after diagnosis,this paper proposes a hybrid attribute information rough set and three-way decision,which expands the application scope of classical rough set,and establishes the corresponding model by integrating machine learning,recommendation technology and multi-attribute decision-making.It also realizes intelligent assistance and scientific decisionmaking in the whole process of diagnosis,and promote the transformation and upgrading of medical service mode.