تحليل خوشه اي مشتريان بر مبناي مدل WRFM با رويكرد داده كاوي غيرنظارتي ( مورد مطالعه محصولات بهداشتي و آرايشي)

Customers Clustring Analysis Based on WRFM Model Using Non-Supervisory Data Mining Approach (Case study of hygienic and cosmetic products)

د ر دﻧﯿﺎي رﻗ ﺎﺑﺘﯽ اﻣﺮو ز ﮐﻪ ﺷﺮﮐﺖ ﻫﺎ ﺑﺎ ﺣﺠﻢ اﻧﺒﻮﻫﯽ از اﻃﻼﻋﺎت ﻣﺸﺘﺮﯾﺎن ﺑﻪ ﻋﻠﺖ رﺷﺪ و ﭘﯿﺸﺮﻓﺖ ﻓﻨﺎوري ﻫﺎي اﻃﻼﻋﺎﺗﯽ و اﯾﺠﺎد ﭘﺎﯾﮕﺎه ﻫﺎي داده اي ﻣﺨﺘﻠﻒ ﻣﻮاﺟﻪ اﻧﺪ اﺳﺘﻔﺎده از اﺑﺰار ﻣﺪﯾﺮﯾﺖ ار ﺗ ﺒﺎط ﺑﺎ ﻣﺸﺘﺮي ﮐﻪ ﺑﺘﻮاﻧﺪ ﺑـﻪ درﺳﺘﯽ و ﺑﻪ ﻣﻮﻗﻊ ﻧﯿﺎزﻫﺎ و اﻧﺘﻈﺎرات ﻣﺸﺘﺮﯾﺎن را ﺷﻨﺎﺳﺎﯾﯽ و رﺻﺪ ﮐﻨﺪ ﺑﯿﺶ از ﭘﯿﺶ ﺿﺮورت ﻣﯽ ﯾﺎﺑﺪ؛ ﯾﮑﯽ از ﺗﮑﻨﯿﮏ ﻫﺎﯾﯽ ﮐﻪ ﻣﯽ ﺗﻮاﻧﺪ در اﯾﻦ ﺑﺮﻫﻪ ﻧﻘﺸﯽ ﮐﻠﯿﺪي و اﺳﺎﺳﯽ اﯾﻔ ﺎﮐﻨﺪ داده ﮐـﺎوي ﭘﺎﯾﮕـﺎه داده ﻫﺎﺳـﺖ . ﻫـﺪف اﯾـﻦ ﭘﮋوﻫﺶ ﺗﺤﻠﯿﻞ داده ﻫﺎي ﻣﺸﺘﺮﯾﺎن ﺑﺮ ﭘﺎﯾﻪ ﻣﺪل WRFM ﺑـﻪ ﮐﻤـﮏ روش ﻫـﺎي داده ﮐـﺎوي ﻏﯿﺮ ﻧﻈـﺎرﺗﯽ اﺳـﺖ ؛ ﭘﮋوﻫﺸﮕﺮ ان درﺻﺪدﻧﺪ ﮐﻪ ﺑ ﺎ ﮐﺸﻒ اﻟﮕﻮﻫﺎي ﻣﻮﺟﻮد ﺑﻪ اراﯾـﻪ اﺳـﺘﺮاﺗﮋي ﻫـﺎي ﻣـﺆﺛﺮﺗﺮي ﺑـﺮاي ﻫـﺮ ﮔـﺮوه از ﻣﺸﺘﺮﯾﺎن و ﺧﺼﻮﺻﺎً ﻣﺸﺘﺮﯾﺎن ﮐﻠﯿﺪي ﺑﭙﺮداز ﻧ ﺪ ﺗ ﺎ ﺑﺮاي ﺳﺎزﻣﺎن ﺳﻮدآوري و ﻋﻤﻠﮑﺮد ﺑﻬﺘﺮي در ﺑﺮداﺷﺘﻪ ﺑﺎﺷﺪ ﺟﺎﻣﻌﻪ آﻣﺎري اﯾﻦ ﭘﮋوﻫﺶ ﻣﺸﺘﺮﯾﺎن ﻣﺤﺼﻮﻻت ﺑﻬﺪاﺷﺘﯽ در ﺑﺎزه زﻣﺎﻧﯽ ﺳﺎل ﻫﺎي 1398-1397اﺳـﺖ ﮐـﻪ ﺑـﻪ روش ﻧﻤﻮﻧﻪ ﮔﯿﺮي ﻫﺪﻓﻤﻨﺪ در دﺳﺘﺮس ﺗﻌﺪاد 64858 ﻧﻤﻮﻧﻪ از ﭘﺎﯾﮕﺎه داده ﻣﺸﺘﺮﯾﺎن اﻧﺘﺨﺎب ﺷﺪه اﺳﺖ. ﺑﻪ ﮐﻤﮏ 3 ﺗﻦ از ﺧﺒﺮﮔﺎن ﻣﺪﯾﺮان ارﺷﺪ ﻓﺮوش ﺷﺮﮐﺖ وزن ﺷﺎﺧﺺ ﻫﺎي ﻣﺪل WRFM ﺗﻌﯿﯿﻦ ﺷﺪه اﺳ ﺖ ﺑﺮاي ﺗﺠﺰﯾـﻪ وﺗﺤﻠﯿﻞ داده ﻫﺎ از ﻧﺮم اﻓﺰار ﮐﻠﻤﻨﺘﺎﯾﻦ و SPSSاﺳﺘﻔﺎده ﺷﺪه اﺳﺖ . ﺑ ﺎ ﺗﻮﺟﻪ ﺑﻪ ﻣـﺪل ﭘـﮋوﻫﺶ ﺧﺎص و ﮐﻠﯿﺪي، ﻃﻼﯾﯽ ﺑﺎﻟﻘﻮه ﻧﺎﻣﻄﻤﺌﻦ ازدﺳﺖ رﻓﺘﻪ، ﻧﺎﻣﻄﻤ ﺌﻦ ﺟﺪﯾﺪ ﺷﻨﺎﺳﺎﯾﯽ و ﻧﺎﻣﮕﺬاري ﺷﺪﻧﺪ ﮐﻪ ﺑﺮاي ﻫﺮ ﯾﮏ از اﯾﻦ دﺳﺘﻪ ﻫ ﺎ اﺳﺘﺮاﺗﮋ ي ﻫﺎي ﻣﺘﻔﺎوﺗﯽ اراﯾﻪ ﺷﺪه اﺳﺖ ﺿﻤﻨﺎ ﻧﺘﺎﯾﺞ ﻧﺸﺎن ﻣﯽ دﻫﺪ ﮐﻪ ﺧﻮﺷﻪ ﺑﻨﺪي K- ﻣﯿـﺎﻧﮕﯿﻦ ﺷﺶ ﺧﻮﺷﻪ اي ﺑﺎ ﺧﻠﻮص 0/744درﺻﺪ ﻧﺴﺒﺖ ﺑﻪ دﯾﮕﺮ روش ﻫﺎ ﻋﻤﻠﮑﺮد ﺑﻬﺘﺮي داﺷﺘﻪ اﺳﺖ

In today's competitive world where companies are faced with a huge amount of customer information due to the growth and development of information technology and the creation of various databases, the use of customer relationship management tools that can accurately and timely identify and monitor customer needs and expectations Becomes more necessary;one of the techniques that can play a key and fundamental role in this period along with this important category is data mining of customer databases. The purpose of this study is to analyze customers clustering based on the WRFM model using non-supervisory data mining methods;the researchers seek to discover the existing rules and patterns to provide more effective strategies for each group of customers, especially key customers, in order to have a better profitability and performance for the organization.Using available purposive sampling method, 64858 samples have been selected from the database of customers who have used hygienic and cosmetic products in the period of 2018-2019.The weight of WRFM attributes has been determined by surveying 3 sales experts (senior managers) of the company. Clementine and SPSS soft wares were used for data analysis. According to the research model, 4 customer categories: Specific and key, Potential golden, lost Uncertainty, New Uncertain were identified and named, and different strategies have been presented for each of these customer categories Also the result show that K-mean clustering with six clusters and purity of 0.744% had better performance than other clustering methods.