Package recommendation method and system

Abstract

The invention relates to the technical field of operators, and particularly provides a package recommendation method and system, and the method comprises the following steps: S1, carrying out the training of a personalized recommendation model; s2, establishing a user satisfaction evaluation model; and S3, personalized recommendation. Compared with the prior art, the user satisfaction and loyalty can be improved, and the competitiveness and sales efficiency of operators can be improved.

Description

Technical Field

[0001] The present invention relates to the field of operator technology, and specifically provides a package recommendation method and system. Background Art

[0002] With the rapid development of the Internet, especially social networks, we are in an era of information overload. It is difficult for users to find content that they are really interested in when faced with excessive information, and it is also difficult for content providers to accurately push high-quality content to interested users. Recommendation systems are considered to be an effective way to solve these problems. They mine users' historical behaviors, model user interests, and predict users' future behaviors, thereby establishing a relationship between users and content.

[0003] Traditional product recommendation methods usually generate recommended products for users based on static rules and predetermined thresholds, lacking a deep understanding of user needs and preferences. In addition, traditional recommendation methods usually have complex logic and configuration options, which require a certain amount of learning cost. Summary of the invention

[0004] The present invention aims at the deficiencies of the above-mentioned prior art and provides a meal plan recommendation method with strong practicality.

[0005] A further technical task of the present invention is to provide a package recommendation system that is reasonably designed, safe and applicable.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] A package recommendation method includes the following steps:

[0008] S1, personalized recommendation model training;

[0009] S2, user satisfaction evaluation model;

[0010] S3. Personalized recommendations.

[0011] Furthermore, in step S1, data models of various businesses are predefined, characteristic information of different businesses is stored in a library, user historical behavior data is pre-imported and processed, RFM and Apache Mahout open source library are used for model training, and user classification models and personalized business recommendation models are generated.

[0012] Furthermore, in step S2, the communication content between the user and the customer service is recorded in the form of voice or text through an online chat window, customer service call records or business hall recordings, and natural language processing technology is used to remove invalid information and obtain key information, extract key feature information, and combine user ratings and evaluations to generate a user satisfaction evaluation model. The model output results are used as the input of the personalized recommendation training model to evaluate and optimize the personalized recommendation model.

[0013] Furthermore, in step S3, according to the analysis results of the personalized recommendation model, specific parameters of the services required by the customers are determined according to the needs of the customers at different stages, the parameters are instantiated, and the recommended products are output by combining the business model and the product model.

[0014] A package recommendation system first trains a personalized recommendation model, then conducts a user satisfaction evaluation model, and uses the model output result as the input of the personalized recommendation training model to evaluate and optimize the personalized recommendation model, and finally conducts personalized recommendation.

[0015] Furthermore, data models for various businesses are pre-defined, feature information of different businesses is stored in the library, user historical behavior data is pre-imported and processed, RFM and Apache Mahout open source library are used for model training, and user classification models and personalized business recommendation models are generated.

[0016] Furthermore, the communication between users and customer service is recorded in voice or text through online chat windows, customer service call records or business hall recordings, and natural language processing technology is used to remove invalid information and obtain key information, extract key feature information, and combine user ratings and evaluations to generate a user satisfaction evaluation model. The model output results are used as the input of the personalized recommendation training model to evaluate and optimize the personalized recommendation model.

[0017] Furthermore, based on the analysis results of the personalized recommendation model, the specific parameters of the services required by customers are determined according to the needs of customers at different stages, the parameters are instantiated, and the recommended products are output by combining the business model and product model.

[0018] Compared with the prior art, the meal plan recommendation method and system of the present invention have the following outstanding beneficial effects:

[0019] The present invention can model customers and identify customer preference behaviors through machine learning based on the user's historical data, and recommend suitable packages to customers; and can analyze customer satisfaction with the recommended content through natural language processing technology based on the text or voice communication content between the customer and the customer service staff, for model re-optimization and model self-learning. Through the analysis of the user's historical one-year consumption behavior, the usage of traffic, voice and text messages within the package, the usage of traffic, voice and text messages outside the package, and other usage behaviors, the specific parameters of the required services can be identified. After identifying the service parameters required by the user, the parameters are instantiated in combination with the product model to generate recommended packages for the user. Ultimately, user satisfaction and loyalty are improved, and the competitiveness and sales efficiency of operators are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Attached Figure 1 This is a schematic diagram of training a personalized recommendation model in a package recommendation method;

[0022] Attached Figure 2 It is a user classification diagram in a package recommendation method;

[0023] Attached Figure 3 This is a schematic diagram of a satisfaction model in a package recommendation method. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention is further described in detail below in conjunction with specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] A best embodiment is given below:

[0026] like Figure 1 , 2 As shown, a package recommendation method in this embodiment has the following steps:

[0027] S1, personalized recommendation model training;

[0028] Pre-define data models for various services, including voice services, traffic services, SMS services, ringback tone, broadband, value-added services, dedicated traffic, night traffic, cloud services, etc., and store characteristic information of different services in the library; pre-import and process user historical behavior data, including consumption behavior, usage of traffic, voice and SMS within the package, usage of traffic, voice and SMS outside the package, recharge amount, recharge cycle, user online time, user registration date, user historical packages, user complaint information, user preferences, user type, etc.

[0029] Based on the above data set, RFM and Apache Mahout open source library are used for model training to generate user classification model and personalized business recommendation model.

[0030] S2, user satisfaction evaluation model;

[0031] like Figure 3 As shown, through online chat windows, customer service call records, or business hall recordings, etc., the content of the communication between users and customer service is recorded in the form of voice or text.

[0032] Use natural language processing technology to remove invalid information and obtain key information, extract key feature information such as business domain, business type, event, evaluation, emotion, satisfaction, etc. and fill in the word slots, combine user ratings and evaluations to generate a user satisfaction evaluation model, and use the model output results as the input of the personalized recommendation training model to evaluate and optimize the personalized recommendation model.

[0033] S3. Personalized recommendations.

[0034] According to the analysis results of the personalized recommendation model, the specific parameters of the services required by customers are determined according to the needs of customers at different stages. The parameters are instantiated, combined with the business model and product model, and recommended products are output.

[0035] Based on the above method, a package recommendation system in this embodiment first trains a personalized recommendation model, then performs a user satisfaction evaluation model, and uses the model output results as input to the personalized recommendation training model to evaluate and optimize the personalized recommendation model, and finally performs personalized recommendation.

[0036] Predefine data models for various businesses, store feature information of different businesses in the library, import and process user historical behavior data in advance, use RFM and Apache Mahout open source library for model training, and generate user classification models and personalized business recommendation models.

[0037] Through online chat windows, customer service call records or business hall recordings, the content of communication between users and customer service is recorded in the form of voice or text, and natural language processing technology is used to remove invalid information and obtain key information, extract key feature information, and combine user ratings and evaluations to generate a user satisfaction evaluation model. The model output results are used as the input of the personalized recommendation training model to evaluate and optimize the personalized recommendation model.

[0038] According to the analysis results of the personalized recommendation model, the specific parameters of the services required by customers are determined according to the needs of customers at different stages. The parameters are instantiated, combined with the business model and product model, and recommended products are output.

[0039] The above-mentioned specific implementations are only specific cases of the present invention. The patent protection scope of the present invention includes but is not limited to the above-mentioned specific implementations. Any technical solutions that comply with the claims of the present invention and any appropriate changes or substitutions made by ordinary technicians in the relevant technical field shall fall within the patent protection scope of the present invention.

[0040] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for recommending a package, characterized in that: The steps are as follows: S1, personalized recommendation model training; S2, user satisfaction evaluation model; S3. Personalized recommendations.

2. A method for recommending a meal plan according to claim 1, characterized in that: In step S1, data models of various businesses are predefined, feature information of different businesses is stored in a library, user historical behavior data is pre-imported and processed, RFM and Apache Mahout open source library are used for model training, and user classification models and personalized business recommendation models are generated.

3. A method for recommending a meal plan according to claim 2, characterized in that: In step S2, the communication between the user and the customer service is recorded in the form of voice or text through the online chat window, customer service call records or business hall recordings, and natural language processing technology is used to remove invalid information and obtain key information, extract key feature information, and combine user ratings to generate a user satisfaction evaluation model. The model output results are used as the input of the personalized recommendation training model to evaluate and optimize the personalized recommendation model.

4. A method for recommending a meal plan according to claim 3, characterized in that: In step S3, according to the analysis results of the personalized recommendation model, the specific parameters of the services required by the customers are determined according to the needs of the customers at different stages, the parameters are instantiated, and the recommended products are output by combining the business model and the product model.

5. A package recommendation system, characterized in that: First, the personalized recommendation model is trained, then the user satisfaction evaluation model is developed, and the model output results are used as the input of the personalized recommendation training model to evaluate and optimize the personalized recommendation model, and finally personalized recommendation is performed.

6. A meal plan recommendation system according to claim 5, characterized in that: Predefine data models for various businesses, store feature information of different businesses in the library, import and process user historical behavior data in advance, use RFM and Apache Mahout open source library for model training, and generate user classification models and personalized business recommendation models.

7. A meal plan recommendation system according to claim 6, characterized in that: Through online chat windows, customer service call records or business hall recordings, the content of communication between users and customer service is recorded in the form of voice or text, and natural language processing technology is used to remove invalid information and obtain key information, extract key feature information, and combine user ratings and evaluations to generate a user satisfaction evaluation model. The model output results are used as the input of the personalized recommendation training model to evaluate and optimize the personalized recommendation model.

8. A meal plan recommendation system according to claim 7, characterized in that: According to the analysis results of the personalized recommendation model, the specific parameters of the services required by customers are determined according to the needs of customers at different stages. The parameters are instantiated, combined with the business model and product model, and recommended products are output.

Images