機器學習應用於客製化餐廳推薦系統

隨著行動裝置的普及化，行動服務逐漸成為商家與消費者溝通及交流的管道，市面上有許多點餐APP，不僅會增加使用者手機的容量，也要花時間了解不同的APP的操作方式，為了解決上述問題，我們設計與開發一套餐廳推薦查詢系統，讓使用者花費較少的時間去找到更適合自己的餐點。本論文設計客製化餐廳推薦資訊系統，使用者透過與LINE機器人的交談方式，將目前所在位置、餐廳種類及預算上傳到系統。根據使用者偏好篩選出符合使用者需求的資料，應用機器學習中線性回歸模型計算出適合使用者的用餐的餐廳，並依照分數高低排序顯示在LINE上供使用者選擇。本系統所使用資料庫包含餐廳資料表、使用者資料表及餐廳評分資料表，餐廳資料表是利用Google Map或手動輸入的餐廳資料建立；使用者資料表由系統註冊來產生；餐廳評分資料表是依據使用者實際消費後，所提供的評分來建立的消費紀錄。此外，資料庫的內容主要針對特定學校周邊餐廳及校內學生消費紀錄，根據實驗結果，可讓該校的學生更容易找出適合自己需求的餐廳。

關鍵字

LINE機器人；機器學習；線性迴歸

並列摘要

With the popularization of mobile devices, mobile services have gradually become a conduit for communication and communication between merchants and consumers. There are many reservation meal apps on the market, which not only increases the capacity of the user's mobile phone, but also takes time to understand how different APPs operate. In order to solve this problem, we design and develop a restaurant recommendation inquiry system, so that users spend less time to find a meal that is more suitable for them. This paper designs a customized restaurant recommendation information system. The user uploads the current location, restaurant type and budget to the system through conversation with the LINE robot. According to the user's preference, the information that meets the user's needs is filtered, and the linear regression model in machine learning is used to calculate the restaurant suitable for the user's meal, and is displayed on the LINE according to the score of the score for the user to select. The database used in this system includes a restaurant data table, a user data table and a restaurant rating data table. The restaurant data table is created by using Google Map or manually input restaurant data; the user data table is generated by system registration, and the restaurant rating data is generated. The table is a consumption record established based on the score provided by the user after actual consumption. In addition, the content of the database is mainly for the consumption records of the surrounding schools and students in the school. According to the experimental results, it is easier for the students to find the restaurant that suits their needs.