The Design And Implementation Of Decision Support System Based On Electronic Records

The traditional decision-making model is based on the static data in the database. In this project, to support business extension and changing, the business data stored in the tables of database must be dynamic. And it makes traditional decision-making model useless for such electronic records. With lots of knowledge behind the electronic records, it is essential to develop a decision support system based on dynamical data in electronic healthy records. It helps to extract data from electronic records and to mine the interesting patterns to implement it in the decision-making. It has the significance to early predict, prevent and diagnose the diseases to raise the people's healthy level and the medical treatment in our country.Taking the problems of dynamic property of data items in electronic records and mining model should be properly selected for different kind of data into account. It is necessary to adopt the strategy of extracting data and constructing mining model instantly, so as to extract data, do data mining and support decision-making dynamically. According to metadata structure of electronic records, the strategy, extracting the data-item from electronic records dynamically and switching the grouped data-item to nested mining structure, is adopted to implement mining models which have the abilities of fault-tolerance and support complex data type.In this project, we create a decision-making platform based on electronic records, and it can select proper mining model for decision-making through following steps: firstly, create temporal tables to store the data of electronic records, and found mining models by some static tables which are consist of temporal tables, then use Microsoft SQL Server 2005 to do decision-making and two exchangeable decision models, neural network and decision tree, are also used, finally put the patterns into knowledge-base and rules-base and set up a model selectable and instant mining decision support platform which is based on electronic records.The platform has been applied by a national 863 project: a medical digital decision support system based on grid. Through experiments of many data, for instance, hypertension data, it shows that this platform has high precision for forecasting disease, and it can provide services for medical staffs, such as exact clinical diagnosis and decision support treatment.