200947329 九、發明說明： 【發明所屬之技術領域】 本發明係關於一種網路購物之系統，尤指一種適用於 網路交易之電子商務網站購物之個人化推薦之系統及其方 5 法。 【先前技術】 〇 如圖1所示，現今的網路購物的網站上都會提供使用者 能夠瀏覽網站的網頁9，或者是搜尋使用者想要尋找的東 10西，虽使用者有需求時並提供購買的功能，使用者能夠不 用外出即可輕輕鬆鬆的在網路上購買所需要的物品或者書 籍，在提供使用者購買的功能時，還會有使用者之購物車 91，以清楚的顯示出使用者目前所想要購買的清單裡有那 坠商m，讓使用者能夠了解自己目前的購物狀況，但目前 15 ^渗】覽、搜尋、購買以及講物車的功能都只有提供使用者 ❹丨純的使用行為而已’並沒有利用這些行為對使用者提供 i動式的推薦服務’而無法提供使用者有效之購物方式， 故其實有予以改善之必要。 20 【發明内容】 一 本發明為一種電子商務網站購物之個人化推薦之系統 及其方法，其系統包括有一用以儲存會員各種行為之記錄 並分析各個會員之關聯性的會員關聯分析模組、一用以儲 存被會員所執行之商品記錄並分析各個商品之關聯性之商 200947329 品 品關聯分析模組、以及―用 用將會員關聯刀析模組及商而 ❹ 10 關聯刀析模組之結果整合出個人商品推薦之系統整合模 組，利用其系統結構’並且先儲存各個會員的所有行為以 及點選過的商品之記錄’其系統之會員關聯分析模組利用 儲存之資料將各個會員的所有行為以及點選過的商品相互 比對’找出關聯性高的其他會員，以及商品關聯分析模組 將各個被會員點選過的商品記錄相互比對，找出商品記錄 關聯性雨的其他商品，最後，將其關聯性高的其他會員以 及其他商品傳送給該系統整合模組以整合結果後推薦給會 員，利用此系統及方法以達成能夠依據會員間之相似行為 模式及喜好來推薦商品，使會員能夠不需經由網路上長時 間的搜尋及瀏覽就能購買可能會有興趣的商品。 【實施方式】 15 本發明為一種電子商務網站購物之個人化推薦之系統 及其方法’如圖2(A)所示，為本發明之系統架構圖，其網 路系統中包括有會員關聯分析模組11、商品關聯分析模組 12、以及系統整合模組13，系統係提供多數種商品，例如 圖書、文具、禮品等，以供會員透過網際網路登入系統來 2〇 購買商品。 請一併參照圖2(B)所示為本發明之電子商務網站購物 之個人化推薦的流程圖，其中該會員關聯分析模組11係用 以儲存會員搜尋、瀏覽、購物、以及購物車之各種行為的 記錄（S100)，並且比對個會員間的各種行為記錄’並且利用 6 200947329 此行為記錄比對的結果來分析出關聯性高的其他會員 (S101)’該商品關聯分析模組12用以儲存被會員所執行之搜 尋、劉覽、、購物、以及購物車的商品記錄，並且比對此商 σ〇 β己錄並且利用此商品記錄之結果來分析出關聯性高的 5 10 15 20 其他商品⑻叫該线整合模組13用以將會員關聯分析模 組及商品關聯分析模組之結果整合出適合個人的商品推薦 (S103)。 圖3(A)所示為會員關聯分析模組u運作之示意圖，一 甲會員和-乙會員在網頁上進行搜尋、㈣、購買商品之 行為以及購物車之商品都被記錄著，系統會將兩個會員的 行為模式記錄進行相互比對，在比對中確認是否有相同的 商品，以及計算出兩個人之間相㈤的商品數量，依據數量 的多寡給予Μ等級’例如圖3(B)所示為會員關聯分析模 組11之一實施例’甲會員的記錄中購買了商品a、b、c， 收藏了商品D、E、F，以及瀏覽了商品G，H，在乙會員記 錄中購W 了商品C、W，記錄著下次再買的商品E、K、L， 所以商品比對的結果顯示’甲與乙共有兩個相同的商品c、 E，則甲與乙之間的關聯性會註記為2 ,再進一步的比對行 為模式和商品的關係，可以發現曱與乙的購買記錄中，共 有-個相同的商品’則甲與乙之間的購買模式關聯性會註 記為卜因此，將每日每個會員的消費、瀏覽、收藏、購物 車、下次再買、wishlist、搜尋等在網站上的一切行為記錄 下二統整後，進行會員與會員間的資料比對，每當會員消 費行為有同—商品時’則增加—點數，最後以點數的多寡 7 200947329 判斷會員之間關聯的強弱，以此強弱度當作權重排序，以 找到跟他行為模式最類似的人，又因為會員的行為會隨時 變動’所以每日進行一次比對’以便更新會員間關聯。 5 ❹ 10 15 ❹ 20 如圖4(A)所示為商品關聯分析模組12運作之示意圖， 一A商品和B商品被會員所執行之搜尋、劉覽、購買之行為 以及放置在購物車將被系統所記錄，並且將兩個商品進行 比對，檢查是否有相同的會員對A商品和8商品執行過相同 的行為，以及計算相同會員的數量，並給序關聯等級，如 圖4(B)所示為商品關聯分析模組12之一實施例，如a商品曾 經被曱和乙會員所購買，以及B商品曾經被甲、乙、丙會員 所購買，其比對結果為A商品和B商品都有被甲、乙會員購 貝，則A商品與b商品之關聯性會註記為2，若假設2為高關 聯性等級，則當丙會員購買B商品後，系統會將八商品推薦 給丙會員。因此，將每個商品被購買、收藏、購物車、下 次再買' wish list、搜尋記錄下來，統整後，進行商品與商 品間的資料比對，每當商品被同一會員執行行為時，則增 加一點數，最後以點數的多募判斷商品之間關聯的強弱， 以此強弱度當作權重排序，使得每個商品能夠找出應會員 行為模式而產生的性質，又因為網站上的行為會隨時變 動’所以每日進行一次比對，以便更新商品間關聯。 如圖5所示為系統整合模組13運作之示意圖，首先，將 曱會員的商品記錄’透過上述的會員關聯分析模組丨丨和商 品關聯分析模組12檢視出與曱會員其關聯性高的會員以及 與曱的商品記錄關聯性高的其他商品，若與曱會員其關聯 8 200947329 5 10 15 20 性馬的會員有丙會員和丁會員，則將丙會員和丁會員的商 品記錄做比對，以找出共同有的商品C、D、E，此時共同 有的商品再和曱的商品記錄綠c、F比對，並且濾掉甲已有 的商品’因此商品D' E為曱所沒有的；在與曱的商品記錄 關聯性高的其他商品部分，找出購買c商品後，大部份會員 的打為為在購買B和D商品以及購買F商品後，會再購買£商 品，因此，整合依據與甲會員其關聯性高的會員和與甲的 商品記錄關聯性高的其他商品所比對的結果，因此，會為 曱會員推薦商品B、D、E’當甲會員收到推薦商品所執 的瀏覽、購買、收藏、沒興趣等等的行為會回饋至系統， 以修正甲會員的行為模式，以達到系統能夠達更準確 薦會員所需要商品。 上述實施例僅係為了方便說明而舉例而已，本 主張之權利範圍自應以中請專利範圍所 於上述實施例。 平而非僅限 【圖式簡單說明】 的示意圖 圖1係習知網路購物的網站所提供之網頁 圖2(A)係本發明之系統架構圖。 圖2(B)係本發明之流程圓。 圖3⑷係本發明之為會員„分析触之示 圓3(B)係會員關聯分析模組之—實施例。 圖4⑷係本發明之為商品與商品間的關聯分析之 圖4(B)係商品與商品間的關聯分析之一實施例。、意圖 9 200947329 圖5係系統整合模組之示意圖。 【主要元件符號說明】 關聯分析模組11 商品關聯分析模組12 系統整合模組13 網頁9 © 購物車91200947329 IX. Description of the Invention: [Technical Field of the Invention] The present invention relates to a system for online shopping, and more particularly to a system for personalization recommendation for e-commerce website shopping for online transactions and a method thereof. [Prior Art] As shown in Figure 1, today's online shopping website provides a web page 9 where users can browse the website, or search for the East 10 West that the user wants to find, although the user has Providing a purchase function, the user can easily purchase the required items or books on the Internet without going out, and when the user purchases the function, the user's shopping cart 91 is displayed for clear display. The user who wants to buy the list now has the merchant's m, so that the user can understand his current shopping situation, but the current function of viewing, searching, purchasing and the parade is only available to the user. The pure use behavior has not used these behaviors to provide users with i-type recommendation services, and it is unable to provide users with effective shopping methods, so it is necessary to improve them. 20 [Summary of the Invention] A system for personalization recommendation of e-commerce website shopping and a method thereof, the system comprising a member association analysis module for storing records of various behaviors of members and analyzing the relevance of each member, A commodity correlation analysis module for storing the product records executed by the members and analyzing the relevance of each product, and the use of the member-associated analysis module and the business-related As a result, the system integration module for personal product recommendation is integrated, and its system structure is used, and all the behaviors of each member and the records of the selected products are stored first. The member association analysis module of the system uses the stored data to All the behaviors and the selected items are compared with each other's other members who have high relevance, and the product association analysis module compares the product records selected by the members to each other to find out the other relevant records of the product records. Commodities, and finally, transfer other members with high relevance and other products to the system integration module. After the results are combined, they are recommended to members. This system and method can be used to recommend products based on similar behavior patterns and preferences among members, so that members can purchase without the need for long-term search and browsing on the Internet. commodity. [Embodiment] 15 The present invention is a system and method for personalization recommendation of e-commerce website shopping, as shown in FIG. 2(A), which is a system architecture diagram of the present invention, and the network system includes member association analysis. The module 11, the product association analysis module 12, and the system integration module 13 provide a variety of products, such as books, stationery, gifts, etc., for members to purchase goods through the Internet login system. Please refer to FIG. 2(B) as a flowchart of the personalized recommendation of the e-commerce website shopping of the present invention, wherein the member association analysis module 11 is used for storing member search, browsing, shopping, and shopping carts. Recording of various behaviors (S100), and comparing the various behavior records between the members' and using 6 200947329 to record the results of the comparison to analyze other members with high relevance (S101) 'The commodity association analysis module 12 It is used to store the merchandise records of the search, the Liu, the shopping, and the shopping cart performed by the member, and analyzes the highly correlated 5 10 15 than the results recorded by the supplier. 20 Other products (8) are called the line integration module 13 for integrating the results of the member association analysis module and the product association analysis module into a product recommendation suitable for an individual (S103). Figure 3 (A) shows a schematic diagram of the operation of the member association analysis module u. One member and one member B search on the web page, (4), the behavior of purchasing the product, and the goods of the shopping cart are recorded, and the system will record The behavior patterns of the two members are compared with each other, and it is confirmed whether there is the same commodity in the comparison, and the number of commodities of the phase (five) between the two persons is calculated, and the Μ grade is given according to the quantity of the quantity', for example, FIG. 3 (B) Shown as an example of the member association analysis module 11 'A member's record purchased the goods a, b, c, collected the goods D, E, F, and browsed the goods G, H, in the B member record In the purchase of goods C, W, recorded the next purchase of goods E, K, L, so the result of the product comparison shows that A and B have two identical goods c, E, then between A and B The relevance of the association will be 2, further comparison of the behavioral patterns and the relationship between the goods, you can find that in the purchase record of 曱 and B, there are a total of the same goods, then the purchase mode association between A and B will be noted For the sake of this, the daily consumption of each member Browse, collect, shop, next buy, wishlist, search, etc. After all the behaviors on the website are recorded, compare the information between the members and the members. When the members consume the same behavior - the goods' Then increase - the number of points, and finally determine the strength of the association between the members by the number of points 7 200947329, using this strength as a weight sort to find the person most similar to his behavior pattern, and because the member's behavior will change at any time 'So make a comparison once a day' to update the membership relationship. 5 ❹ 10 15 ❹ 20 As shown in Fig. 4(A), the operation of the product association analysis module 12 is shown. The search for a product and the B product by the member, the tour, the purchase behavior, and the placement in the shopping cart will be It is recorded by the system, and the two commodities are compared, and it is checked whether the same member has performed the same behavior for the A commodity and the 8 commodity, and the number of the same member is calculated, and the order of the association is given, as shown in FIG. 4 (B). Shown as an embodiment of the product association analysis module 12, such as a product has been purchased by the 曱 and B members, and B products have been purchased by the A, B, C members, the comparison results are A goods and B If the goods are purchased by A and B members, the relevance of A products to b products will be recorded as 2, and if 2 is a high relevance level, then when C members purchase B products, the system will recommend eight products. C member. Therefore, each item is purchased, collected, shopping cart, and next time you buy 'wish list, search and record, and then compare the data between the goods and the goods. When the goods are executed by the same member, Then increase the number of points, and finally judge the strength of the relationship between the products by the number of points, and use this strength as the weight order, so that each product can find out the nature of the member's behavior pattern, and because of the website The behavior will change at any time' so a daily comparison is made to update the association between the items. FIG. 5 is a schematic diagram showing the operation of the system integration module 13. First, the product record of the member is viewed through the member association analysis module and the product association analysis module 12 described above, and the association with the member is high. Members and other products that are highly relevant to 商品's product records, if they are associated with 曱 members, 8 200947329 5 10 15 20 Sexual horse members have C members and Ding members, then the C members and Ding members' product records are compared In order to find out the common products C, D, and E, at this time, the common products are compared with the product records of the green, c, and F, and the existing products are filtered out. Therefore, the product D'E is 曱If there is no other product part with high relevance to the product record, after finding out the purchase of the c product, most of the members will purchase the product after purchasing the B and D products and purchasing the F product. Therefore, the integration is based on the results of comparisons between members who have high relevance to member A and other products that are highly relevant to A's product records. Therefore, products B, D, and E' are recommended for members. To the recommended goods Browse, purchase, collection, etc. are not interested in the behavior of the system will be back to order correction A behavior patterns of members, in order to achieve a more accurate recommendation system capable of members needed goods. The above-described embodiments are merely examples for the convenience of the description, and the scope of the claims is the scope of the above-mentioned embodiments. Figure 1 (A) is a system architecture diagram of the present invention. Figure 2 (A) is a system architecture diagram of the present invention. Figure 2 (B) is the flow circle of the present invention. Fig. 3 (4) is an embodiment of the present invention as an analysis of the member of the analysis of the circle 3 (B) member association analysis module. Fig. 4 (4) is a diagram of the correlation analysis between the commodity and the commodity of the present invention. An example of the correlation analysis between goods and goods. Intention 9 200947329 Figure 5 is a schematic diagram of the system integration module. [Key component symbol description] Association analysis module 11 Commodity correlation analysis module 12 System integration module 13 Web page 9 © Shopping Cart 91