چكيده به لاتين
Abstract:
In this research, by means of a mixed framework of Data Mining and Data Envelopment Analysis, a new procedure is provided for relative evaluation of the hospitals of province Tehran, in which the inputs of the hospitals are the accreditation factors of each part of the hospitals that in which, instead of just the quantity of each input, the quality of which is indicated in evaluation system. In this research, betwin 25 accreditaion scores of different parts of hospitals, 7 factors of them have been chosen that with 9 outputs of hospitals, have been used in Data Envelopment Analysis process. In order to evaluate, DEA-CCR input oriented method has been employed to do the relative evaluation process. In this evaluation the undesired outputs have seen in the DEA process, which are the number of death occurance before and after 24 hours and the average of hospitalization days. At the end, by means of shadow variables gained from DEA model, a virtual pattern has been extracted for each deficient hospital that each deficient hospital can reach its own maximum efficiency by put that pattern as its goal. The most important factors which have been recognized through this research and been used as inputs for in DEA process, were management and leadership, infection control, nursing management, information technology, human resource management and drug management. At the end, the Total Accreditaion index, which is used for rating the hospitals, and the DEA index have been compared against each other that which one can be more accurate in prediction of hospitals ability, as the result, the DEA index have been recognized as a more accurate index in estimate the hospitals ability.
Keywords: Hospital evaluation, Accreditaion, Data envelopment analysis, Datamining, Dimention reduction.
