Self-service ordering recommendation system and method thereof

Abstract

A self-service ordering recommendation system is disclosed by the present disclosure. The self-service ordering recommendation system comprises a restaurant search module, a review text acquisition module, a natural language processing module, a scoring calculation module, a meal recommendation module and a system integration module. A score is assigned to each one of dining experience keywords by the scoring calculation module according to a score of a corresponding positive keyword or a score of a negative keyword, and the scores of the dining experience keywords are calculated, thereby generating a scoring result of a plurality of meals of a restaurant, so as to obtain a meal positive scoring list from the scoring result of the plurality of meals as a reference for a user for a selection of a meal. A self-service meal ordering recommendation method is further provided by the present disclosure.

Claims

1. A self-service ordering and recommendation system, comprising: a restaurant search module (10) for providing a user with the ability to search the Internet by entering the name of a restaurant as a keyword to find websites with reviews of the restaurant; a review text extraction module (20) for extracting the text of reviews of the restaurant on the Internet and storing the text of the reviews as data to be analyzed, the data to be analyzed including the restaurant name, the name of the dish and the dish review; a natural language processing module (30) for performing natural language processing and sentiment analysis on the dish review in the data to be analyzed to obtain a dining experience keyword related to the dining experience, and classifying the dining experience keyword into a positive review keyword group or a negative review keyword group to obtain a classification result, wherein the positive review keyword group includes a plurality of positive keywords and each positive keyword has a corresponding score, and the negative review keyword group includes a plurality of negative keywords and each negative keyword has a corresponding score; A rating calculation module (40) is used to assign a score to each of the eating experience keywords based on the positive review keyword group or the negative review keyword group, and to count the scores of multiple eating experience keywords to obtain a rating result for the meal, thereby obtaining the rating results of multiple meals in the restaurant, so as to select specific meals from multiple meals in the restaurant based on the rating results of multiple meals to generate a positive rating list for a meal; A restaurant recommendation module (50) is used to provide a positive rating list of the restaurant to a user; and a system integration module (60) is electrically connected to the restaurant search module (10), the review text extraction module (20), the natural language processing module (30), the rating calculation module (40) and the restaurant recommendation module (50) respectively, and is used to perform the operation of the restaurant search module (10), the review text extraction module (20), the natural language processing module (30), the rating calculation module (40) and the restaurant recommendation module (50).

2. The self-service ordering and recommendation system as described in request item 1, wherein, The data to be analyzed also includes restaurant response time, number of restaurant responses, and restaurant response content. Furthermore, the rating calculation module (40) also includes a bonus value calculation module (41) to set a bonus value for the restaurant response time, the number of restaurant responses, and the restaurant response content in the data to be analyzed, so as to obtain a response time bonus value, a response number bonus value, and a response content bonus value. The rating calculation module (40) calculates the scores of multiple dining experience keywords together with one or more of the groups composed of the response time bonus value, the response number bonus value, and the response content bonus value to obtain the rating result of the meal.

3. The self-service ordering and recommendation system as described in request item 1, wherein, The comment text extraction module (20) also includes a language type conversion unit (21), which is electrically connected to the system integration module (60) to identify the language type of the data to be analyzed and to perform language type conversion of the data to be analyzed.

4. The self-service ordering and recommendation system as described in request item 1, wherein, The comment text extraction module (20) also includes a date analysis unit (22) for analyzing the publication date of the restaurant's comments and setting a comment period, which is the number of days from the publication date of the restaurant's comments to the search date. The comment text extraction module (20) only extracts the restaurant's comments within the comment period.

5. A self-service ordering recommendation method using the self-service ordering recommendation system as described in claim 1, comprising: a restaurant search step (S1), using a restaurant search module (10) to provide a user with the ability to search the internet by entering the name of a restaurant as a keyword, thereby finding websites with reviews of the restaurant; a review text extraction step (S2), using a review text extraction module (20) to extract the text of the searched review, and storing the text of the review as data to be analyzed, the data to be analyzed including the restaurant name, the name of the dish, and the dish review; and a natural language analysis step (S3), using a natural language processing module (30). Natural language processing and sentiment analysis are performed on the meal reviews in the data to be analyzed to obtain keywords related to the meal experience. These keywords are then categorized into a positive comment keyword group or a negative comment keyword group to obtain a classification result. The positive comment keyword group contains multiple positive keywords, each with a corresponding score, and the negative comment keyword group contains multiple negative keywords, each with a corresponding score. A rating calculation step (S4) uses a rating calculation module (40) to assign a score to each of the food review keywords based on the positive review keyword group or the negative review keyword group, and to count the scores of multiple food review keywords to obtain a rating result for the meal, thereby obtaining rating results for multiple meals in the restaurant, and selecting specific meals from the multiple meals in the restaurant based on the rating results of the multiple meals to generate a positive rating list for meals; and a meal recommendation step (S5) uses a meal recommendation module (50) to provide the positive rating list for meals to a user.

6. The self-service ordering recommendation method as described in request item 5, wherein, The comment extraction step (S2) further includes a language type conversion step (S21), which uses a language type conversion unit (21) to identify the language type of the data to be analyzed and to perform language type conversion of the data to be analyzed.

7. The self-service ordering recommendation method as described in request item 5, wherein, The comment retrieval step (S2) further includes a comment period analysis step (S22), which uses a date analysis unit (22) to analyze the publication date of the restaurant's comments and sets a comment period, which is the number of days that the restaurant's comments have been in since the publication date to the current search date. The comment text retrieval module (20) only retrieves the restaurant's comments within the comment period.

8. The self-service ordering recommendation method as described in request item 5, wherein, In the restaurant search step (S1), the restaurant search module (10) searches for websites with a specific number of reviews based on a search threshold, wherein the search threshold is set based on the number of reviews on the website.

9. The self-service ordering recommendation method as described in request item 5, wherein, In the rating calculation step (S4), the rating calculation module (40) further ranks the plurality of meals in the positive rating list of the meal according to their rating results; and in the meal recommendation step (S5), the meal recommendation module (50) recommends a specific meal in the positive rating list of the meal according to a popular recommendation threshold, wherein the popular recommendation threshold is the ranking number of the plurality of meals.

10. The self-service ordering recommendation method as described in request item 5, wherein, The data to be analyzed also includes restaurant response time, number of restaurant responses, and restaurant response content. Furthermore, the rating calculation step (S4) includes a bonus value calculation step (S41), and the rating calculation module (40) includes a bonus value calculation module (41). In the bonus value calculation step (S41), the bonus value calculation module (41) sets a bonus value for the restaurant response time, the number of restaurant responses, and the restaurant response content in the data to be analyzed, so as to obtain a response time bonus value, a response number bonus value, and a response content bonus value. The rating calculation module (40) calculates the scores of multiple dining experience keywords together with one or more of the groups composed of the response time bonus value, the response number bonus value, and the response content bonus value to obtain the rating result of the meal.

Images