Here’s the table of contents:

Dish recommendation system based on graph database

     This article is from a community expert pangguoming Shared map database application cases ： Use Neo4j And a simple word segmentation algorithm to implement the dishes recommendation system , The full text is as follows ：

     background : This recommended system is based on a hardware product – Wang Xiaobao table card . The guest presses the order button , Scan the code to enter the order interface , Then start ordering your favorite dishes , Place an order on the mobile terminal . At present, it has nearly 200 Cooperative catering businesses .

     Dish recommendation function :
When a guest orders at a store with a table card , Extract the main ingredients of the guest's order , The next time a guest orders at another shop that uses table cards , That is to recommend the main ingredients of the merchant for the customer .

     Such as ： A guest is A Shop ordered [ Spicy chicken feet ], Then the favorite ingredients of the guest are ” Chicken feet ”, When the guests come B shop , Can recommend B Corresponding to the store [ Er Niang chicken feet ].

One 、 Extraction of main ingredients of dishes

     The name of the dish on the table is entered by the merchant , Save to the table database , There is an appointment after the name of the existing dish is removed 2 m .

• First step : Collect dish names , Export the dish name from the database .
  
• The second step : participle & Word frequency statistics ： You can use open source word segmentation tools , In this case, we use Word Word segmentation is .
  
• The third step : Manual screening of main materials ： The higher the frequency of words , The more frequently they appear in the dish names , The more valuable it is in screening ; The frequency of words is 1 There is no need to filter the words , Because even the main material , There are no other dishes to recommend .
  
• Step four : Match the main material algorithm ： The specific algorithm can be determined by the business scenario , The matching result is as follows ,“=>” On the left is the main material , On the right is the matched dish name .
  

Two 、 data structure

     In this system , It's about “ people - shop - food - Main material ” Relationship , In order to simplify the structure of relationships , So it introduces Neo4j Graphic database , In the graphics database , The relationship is as follows .

     When the guest arrives , Recommend to the guest that the restaurant can match the dishes he likes , Sort by preference weight .

3、 ... and 、 System architecture