The Application Of Cluster Analysis Algorithm In HMIS

As the informationization level of ascension, many domestic hospitals have set uphospital information management system. Concrete can be divided into again, be in hospitaloutpatient service system, the electronic medical record system, examination system, drugmanagement system, nursing system, financial management system, equipment managementsystem, query system and support system maintenance, etc. Required to run the system tostore a lot of hospital about the data, how can you conclude from these systems in the datathat managers need information, is not a easy to achieve.Hospital management system in the query and management of finance, medicine, iswidely used in hospital management, and only the general data processing, the data is realdeep utilization is less. Therefore ascension information construction in hospital at the sametime, also need to realize mining association use and deep analysis of the data, so as to realizethe data reprocessing, and application, make the data to repetition, make full use ofinformation resources, and get the gain.This article is based on the HMIS system database, combined with the historical data ofthe hospital, and long-term accumulated report, report and other data, according to the specialneeds of the business, hospital has established departments such a topic, efficiency andbenefit management by using clustering analysis in data mining algorithm, based on the sevenindicators analyzing the status quo of the18major clinical departments, and for managers tomake targeted solution Suggestions.In this paper, the main work done:A. In-depth study of data mining technology.1, the data preprocessingHMIS database of large amount of data, the data source is more, a lot of low quality ofdata, such as noise, missing, inconsistent, can lead to the failure analysis results of datamining. Data preprocessing is to raise the quality of the source data, the data mining processmore effective and more convenient. Data preprocessing of main steps: data cleaning, dataintegration, data reduction and data transformation.2, research and the establishment of data warehouse Data warehouse is the data extraction operation system and as a historical snapshot queryand is scheduled to report immediately, to organize the data warehouse according to the datacharacteristics of the data. Data warehouse model creation process: configurationmanagement, data warehouse, theme design, data preprocessing, multidimensional datadimension design, design of fact table and dimension table, and finally to create a hospital inthe data warehouse model.3, clustering algorithmClustering can be considered to be one of the most important problem of unsupervisedlearning, so it involves looking for a collection of unlabeled data structure. A cluster is acollection of objects, used to have similar characteristics in the collection of objects to thesame organization. In this paper, we study the clustering algorithm with hierarchicalclustering and fuzzy clustering and K-Means algorithm.B. Analysis of the application and limitations in hospitalC. data warehouse, build hospital to hospital HMIS system database data and historicaldata base, according to the data mining process in the first place to perform datapreprocessing, data configuration management, the data warehouse theme design, design ofmultidimensional data dimension, the fact table and dimension table design, finally to create ahospital in star structure model of the data warehouse.D. the use of data mining technology of three kind of clustering algorithm, fuzzyclustering and hierarchical clustering and K-Means algorithm, and determines the clusteringanalysis of seven indicators:3patients with the diagnosis rate and cure rate, in-hospital time,department sickbed utilization rate, each patient the average length of hospital stay, clinicalmedicine, annual income proportion. Clustering analysis18hospital clinical departments andget the expected classification.For clustering analysis of the five characteristics of classification, this paper for themanagement of the hospital policymakers put forward the corresponding solution Suggestions,for the scientific management of the hospital and department efficiency, benefit provides apowerful reference.