مدل سازي ارزيابي مشتريان از كيفيت خدمات درماني با استفاده از شبكه عصبي مصنوعي مطالعه موردي بيمارستان هاي دانشگاه علوم پزشكي بيرجند

Modeling Customer Evaluations of the Quality of Health Care Using Artificial Neural Network ( Case Study of Birjand University of Medical Sciences )

مقدمه: كيفيت خدمات همواره يكي از مهمترين دغدغه هاي مديران در تامين رضايت مشتريان است. تامين خدمات با كيفيت نيازمند شناخت دقيق نسبت به عوامل كليدي كيفيت خدمت و نوع اثربخشي آنها در ميزان رضايت مشتريان است. لذا بكارگيري روشهاي مختلف سنجش كيفيت خدمات ميتواند جنبه هاي مجهول اثربخشي اين عامل را بر رضايت مشتريان شفاف تر سازد. از اينرو هدف از انجام اين تحقيق ارزيابي روش هاي سنجش كيفيت خدمات درماني با استفاده از شبكه هاي عصبي مصنوعي بوده است. روش پژوهش: اين تحقيق به روش توصيفي - همبستگي انجام شده و از نوع تحقيقات كاربردي است. جامعه آماري تحقيق را كليه مشتريان بيمارستان هاي دانشگاه علوم پزشكي بيرجند تشكيل مي دهند كه تعداد آنها معين نبوده است. با استناد به رابطه نمونه گيري كوكران تعداد ۳۸۵ نفر از آنها به شيوه در دسترس انتخاب شده و پرسشنامه هاي اعتبارسنجي شده تحقيق ميان آنان توزيع گرديد. به منظور سنجش كيفيت خدمات از 4 روش سروكوآل وزني و غير وزني و سروپرف وزني و غيروزني استفاده شد و تاثير ابعاد كيفيت خدمات در هريك از اين 4 روش بر روي رضايت مشتريان مورد ارزيابي قرار گرفت. در اين تحقيق براي تجزيه و تحليل داده ها ازنرم افزار spss استفاده شده است و نتايج جهار روش سنجش كيفيت خدمات با استفاده از شبكه هاي عصبي مصنوعي مورد بررسي قرار گرفته است. يافته ها: نتايج نشان داد كه از بين روش هاي سنجش كيفيت خدمات كمترين سطح خطا براي مدل سروكوال وزني عدد ۰،۱۸ بدست آمده كه سنجش كيفيت خدمات بر حسب مدل سروكوال وزني با استفاده از شبكه هاي عصبي مصنوعي دقت بالاتري در پيش بيني رضايت مشتريان داشته است. نتيجه گيري: روش هاي سنجش كيفيت خدمات تحت تاثير مقياس سنجش كيفيت خدمات، عملكرد متفاوتي در پيش بيني رضايت مشتريان دارند. همچنين روش هاي شبكه عصبي مصنوعي نسبت به الگوريتم بكارگرفته شده در پيش بيني، ممكن است پيش بيني هايي ضعيف تر از نتايج روش هاي آمار كلاسيك ارائه دهند.

Introduction: The service quality is always one of the managerial concerns to supply customer’s satisfaction. Preparing qualified service needs to exact knowledge about the key factors of service quality and their effectiveness in the level of customer’s satisfaction. So implementing the different methods of measuring service quality could make it more explicit the unknown aspects of this factor effectiveness on the satisfaction. So the aim of this study was to evaluating the health care quality methods with artificial neural network approach. Methods: This study was a descriptive-correlation and an applied research. The statistical population of research consists of customers in hospitals of medical sciences Birjand University with an indefinite number. Referring to Cochran sampling formula a number of 385 individuals were selected using in access approach and validated questionnaires of study distributed among them. To measure the service quality it used the 4 approaches of weighted and un-weighted SERVQL and SERVPRF and the effect of service quality dimensions in each 4 approach were evaluated on the satisfaction. In this study to analyze the data is used of Spss software and the results of four methods to measure service quality using artificial neural networks have been studied. Results: The results showed that the method of measuring the quality of services achieved the lowest level of error for SERVQUAL 0.18 Weighted number That measure the quality of service in terms of weight SERVQUAL model using artificial neural networks have been more accurate in predicting customer satisfaction. Conclusions: methods of measuring service quality have different performance in predicting customer’s satisfaction under the scale of measuring service quality. Also the artificial neural networks regarding to implement predicting algorithm, may contain weaker forecast rather than classic statistical methods. Introduction: uality of service has always been one of the main concerns of managers in providing customer satisfaction. So, employing different methods to measure the effectiveness of this agent's unknown aspects of service quality can be more transparent on customer satisfaction. Methodology: This study was conducted cross-correlation functional investigation. The population of Birjand University of Medical Sciences were all customers that their number was not specified. According to Cochran sampling about 385 of them were selected based on availability of validated questionnaires were distributed among them. To measure the service quality it used the 4 approaches of weighted and un-weighted SERVQL and SERVPRF and the effect of service quality dimensions in each 4 approach were evaluated on the satisfaction. The data were analyzed using multi-layered artificial neural networks.Findings:The results showed that the method of measuring the quality of services achieved the lowest level of error for SERVQUAL 0.18 Weighted number That measure the quality of service in terms of weight SERVQUAL model using artificial neural networks have been more accurate in predicting customer satisfacti Conclusion: methods of measuring service quality have different performance in predicting customer’s satisfaction under the scale of measuring service quality. Also the artificial neural networks regarding to implement predicting algorithm, may contain weaker forecast rather than classic statistical methods