Hotel recommendation method and device based on user behavior analysis

Abstract

The present application discloses a hotel recommendation method and device based on user behavior analysis, and the server method includes: receiving a hotel reservation request sent by a client, the hotel reservation request carries a destination and a target user identifier; obtaining the target travel standard corresponding to the target user identifier; screening hotel data within a preset range around the destination and meeting the target travel standard, and generating a set of candidate hotels; obtaining historical user behavior data corresponding to the target user identifier before the current moment; fitting a factor change curve of multiple preset hotel screening factors of the target user to which the target user identifier belongs within a preset period based on the user behavior data; inputting the factor change curve into a pre-trained personal preference prediction model, and outputting a preference keyword sequence; from the set of candidate hotels, screening hotels that meet the preference keyword sequence and recommending them to the client. Therefore, by adopting the present application, the recommended hotels can meet the user's expectations, thereby improving the accuracy of hotel recommendations.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a hotel recommendation method and device based on user behavior analysis. Background Art

[0002] Business travelers often face tight time constraints and strict travel standards. They need a system that can quickly and accurately recommend hotels that meet their budget and quality requirements. In addition, with the progress of globalization, business people may need to travel frequently between different cities or even countries, which increases the complexity of finding suitable hotels. Therefore, a hotel recommendation system that can adapt to different regions and cultural differences is particularly important for business travelers.

[0003] In the related art, there are currently a variety of online platforms on the market that provide hotel reservation services. These platforms provide users with hotel options by collecting and analyzing basic hotel information, such as location, price, facilities, etc. These platforms often fail to fully consider the personalized needs of users at different time periods, resulting in the recommended hotels not fully meeting the user's expectations, thereby reducing the accuracy of hotel recommendations. Summary of the invention

[0004] The embodiments of the present application provide a hotel recommendation method and device based on user behavior analysis. In order to have a basic understanding of some aspects of the disclosed embodiments, a simple summary is given below. This summary is not a general review, nor is it intended to identify key / important components or describe the scope of protection of these embodiments. Its only purpose is to present some concepts in a simple form as a preface to the detailed description that follows.

[0005] In a first aspect, an embodiment of the present application provides a hotel recommendation method based on user behavior analysis, which is applied to a server, and the method includes:

[0006] Receive a hotel reservation request sent by a client, where the hotel reservation request is used to request the server to generate an optimal hotel list belonging to the user, and the hotel reservation request carries a destination and a target user identifier;

[0007] Obtaining the target travel standard corresponding to the target user identifier from the mapping relationship between the pre-configured user identifier and the user travel standard;

[0008] Filter hotel data within a preset range around the destination that meets the target travel standards to generate a set of candidate hotels;

[0009] From the pre-collected dynamic mapping relationship between user IDs and user behavior information, all historical user behavior data corresponding to the target user ID before the current moment are obtained. The user behavior information in the dynamic mapping relationship is stored one by one in chronological order and obtained from real-time monitoring of the target user ID by a preset monitor;

[0010] Based on the user behavior data, a factor change curve of multiple preset hotel screening factors of the target user to which the target user ID belongs within a preset period is fitted. The preset hotel screening factors are used to describe the basic attributes and conditions of hotel characteristics. The factor change curve is used to characterize the change in the degree of attention paid by the target user to each hotel screening factor at different time points;

[0011] Inputting factor change curves of multiple preset hotel screening factors into a pre-trained personal preference prediction model, outputting a preference keyword sequence corresponding to the target user, the preference keyword sequence is used to represent the user's personalized demand for hotels;

[0012] From the set of candidate hotels, select all hotels that meet the preferred keyword sequence as the optimal hotel list and recommend them to the client for display.

[0013] Optionally, the pre-trained personal preference prediction model includes a region segmentation module, a preference strength extraction module, a context information collection module, a coding fusion module, and a random forest regression network;

[0014] Input the factor change curves of multiple preset hotel screening factors into the pre-trained personal preference prediction model, and output the preference keyword sequence corresponding to the target user, including:

[0015] The region segmentation module segments the region where the hotel screening factor with the highest frequency appears in the factor change curve diagram of multiple preset hotel screening factors as the focus region, and the focus region is used to represent the hotel attributes preferred by the target user in different time periods;

[0016] The preference strength extraction module extracts and quantifies the target user's preference strength for hotel screening factors based on the focus area;

[0017] The context information collection module collects context information related to the target user, and the context information is used to characterize the user's hotel booking background;

[0018] The encoding fusion module encodes and fuses the preference strength and context information to obtain fusion features;

[0019] The fused features and the preset preference keyword set are input into the random forest regression network to output the preference keyword sequence corresponding to the target user.

[0020] Optionally, based on the focus area, extract and quantify the target user's preference intensity for hotel screening factors, including:

[0021] From the preset hotel screening factor set, select the target factor located in the area of ​​interest;

[0022] Collect hotel star ratings, prices, user reviews, and booking frequencies of target users in the area of ​​interest to obtain historical booking data;

[0023] Based on historical booking data, the comprehensive score of the target factor is determined as the target user's preference intensity for the hotel screening factor.

[0024] Optionally, generate a pre-trained personal preference prediction model by following these steps, including:

[0025] Collect historical hotel booking data for each historical user;

[0026] According to the historical hotel booking data of each historical user, a historical factor change curve chart of multiple preset hotel screening factors of each historical user within a preset period is fitted;

[0027] Create a personal preference prediction model, which includes a region segmentation module, a preference strength extraction module, a context information collection module, a coding fusion module, and a random forest regression network;

[0028] The regional segmentation module and the preference strength extraction module are used to process the historical factor change curve graph in turn to obtain the historical preference strength corresponding to each historical factor change curve graph;

[0029] A context information collection module is used to collect historical context information related to each historical user;

[0030] The encoding fusion module is used to fuse the historical preference strength and historical context information corresponding to each historical factor change curve to obtain the historical fusion features of each historical user;

[0031] Based on the historical fusion features of each historical user, the random forest regression network is machine learned to obtain a pre-trained personal preference prediction model.

[0032] Optionally, based on the historical fusion features of each historical user, a random forest regression network is machine-learned to obtain a pre-trained personal preference prediction model, including:

[0033] Initialize the preset favorite keyword set;

[0034] In response to the data annotation instruction, a target preference keyword corresponding to the data annotation instruction is selected from a preset preference keyword set to perform data annotation on the historical fusion features of each historical user to obtain a plurality of model training samples;

[0035] Input each model training sample into the random forest regression network and output the model loss value;

[0036] Based on the model loss value, a pre-trained personal preference prediction model is generated.

[0037] Optionally, generate a pre-trained personal preference prediction model based on the model loss value, including:

[0038] When the model loss value reaches the minimum, a pre-trained personal preference prediction model is generated;

[0039] or,

[0040] When the model loss value has not reached the minimum, the network parameters of the random forest regression network are updated, and the step of inputting each model training sample into the random forest regression network is continued until the loss value reaches the minimum; wherein,

[0041] The loss function of the random forest regression network is:

[0042]

[0043] in, is the model loss value, Indicates the first samples, represents the classification loss, Represents the random forest regression network for the Sample predicted preferred keywords, Expressing the The sample annotated preferred keyword labels, is the weight to balance the loss.

[0044] Optionally, user behavior data includes searched keywords, browsed hotel detail pages, clicked hotel pictures or descriptions;

[0045] Based on the user behavior data, a factor change curve diagram of multiple preset hotel screening factors of the target user to which the target user ID belongs within a preset period is fitted, including:

[0046] Get multiple preset hotel filter factors;

[0047] Determine the preset period;

[0048] Draw a two-dimensional coordinate graph based on a preset period and multiple preset hotel screening factors;

[0049] The searched keywords, the browsed hotel details pages, and the clicked hotel pictures or descriptions are mapped to a two-dimensional coordinate graph to obtain a factor change curve graph of multiple preset hotel screening factors of the target user to which the target user identification belongs within a preset period.

[0050] Optionally, all hotels that meet the preferred keyword sequence are selected from the candidate hotel set as the optimal hotel list, including:

[0051] Obtain hotel information and historical user evaluation information for each candidate hotel in the candidate hotel set;

[0052] Determine whether each preference keyword in the preference keyword sequence exists in the hotel information and historical user evaluation information of each candidate hotel, and obtain a determination result;

[0053] From the determination results, count the number of each preference keyword in the preference keyword sequence in the hotel information of each candidate hotel and the historical user evaluation information;

[0054] Based on the number of each preference keyword in the preference keyword sequence, an optimal hotel list is generated.

[0055] Optionally, based on the number of each preference keyword in the preference keyword sequence, an optimal hotel list is generated, including:

[0056] Acquire hotel information in which the number of each preference keyword in the preference keyword sequence is greater than a preset threshold and candidate hotels corresponding to historical user evaluation information;

[0057] The obtained candidate hotels are arranged in descending order according to the number of preference keywords in the preference keyword sequence to obtain an optimal hotel list.

[0058] In a second aspect, an embodiment of the present application provides a hotel recommendation device based on user behavior analysis, the device comprising:

[0059] A request receiving module, used to receive a hotel reservation request sent by a client, the hotel reservation request is used to request the server to generate an optimal hotel list belonging to the user, and the hotel reservation request carries a destination and a target user identifier;

[0060] A travel standard acquisition module is used to acquire the target travel standard corresponding to the target user identifier from the mapping relationship between the pre-configured user identifier and the user travel standard;

[0061] The hotel data screening module is used to screen hotel data within a preset range around the destination that meets the target travel standards and generate a set of candidate hotels;

[0062] The user behavior data acquisition module is used to obtain all historical user behavior data corresponding to the target user identity before the current moment from the dynamic mapping relationship between the pre-collected user identity and the user behavior information. The user behavior information in the dynamic mapping relationship is stored one by one in chronological order and obtained from the real-time monitoring of the target user identity by the preset monitor;

[0063] A factor change curve fitting module is used to fit a factor change curve of multiple preset hotel screening factors of a target user to which a target user ID belongs within a preset period based on user behavior data. The preset hotel screening factors are used to describe the basic attributes and conditions of hotel characteristics. The factor change curve is used to characterize the changes in the degree of attention paid by the target user to each hotel screening factor at different time points;

[0064] A preference keyword sequence output module is used to input factor change curves of multiple preset hotel screening factors into a pre-trained personal preference prediction model, and output a preference keyword sequence corresponding to the target user, where the preference keyword sequence is used to represent the user's personalized demand for hotels;

[0065] The optimal hotel list screening module is used to screen all hotels that meet the preferred keyword sequence from the candidate hotel set as the optimal hotel list, and recommend them to the client for display.

[0066] The technical solution provided by the embodiments of the present application may have the following beneficial effects:

[0067] In an embodiment of the present application, on the one hand, the server makes hotel recommendations by obtaining the user's travel standards and historical behavior data. The historical behavior data is collected in real time through a preset listener and reflects the user's attention to hotel features at different time points. By analyzing these behavior data, the server can have a deeper understanding of the user's personalized needs and preference changes, so that the recommended hotels not only meet the user's travel standards, but also meet their implicit personal preferences. This method of combining user behavior data improves the accuracy of hotel recommendations; on the other hand, based on user behavior data, the server fits factor change curves of multiple preset hotel screening factors. These curves reveal the changes in the user's attention to different hotel attributes within a preset period. These factor change curves provide rich input features for the personal preference prediction model, so that the model can capture the dynamic changes in user preferences, thereby enhancing the personalization of hotel recommendations.

[0068] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0070] Figure 1 This is a method flow diagram of a hotel recommendation method based on user behavior analysis provided by an embodiment of the present application;

[0071] Figure 2 This is a data representation of a mapping relationship between a user identifier and a travel standard provided in an embodiment of the present application;

[0072] Figure 3 This application provides a data representation of the mapping relationship between user identification and historical user behavior data;

[0073] Figure 4 It is a schematic diagram of the model architecture of a personal preference prediction model provided by this application;

[0074] Figure 5 It is a flowchart of a model training method for a personal preference prediction model provided by the present application;

[0075] Figure 6 This is a schematic diagram of the client interface of an optimal hotel list provided by this application;

[0076] Figure 7 It is a structural schematic diagram of a hotel recommendation device based on user behavior analysis provided by this application;

[0077] Figure 8 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0078] The following description and the accompanying drawings sufficiently illustrate specific embodiments of the present application to enable those skilled in the art to practice them.

[0079] It should be clear that the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0080] When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Instead, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.

[0081] In the description of the present application, it should be understood that the terms "first", "second", etc. are used for descriptive purposes only and should not be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances. In addition, in the description of the present application, unless otherwise specified, "multiple" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects associated before and after are in an "or" relationship.

[0082] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0083] Currently, there are many online platforms in the market that provide hotel reservation services. These platforms provide users with hotel options by collecting and analyzing basic information about hotels, such as location, price, facilities, etc.

[0084] The inventors realize that these existing technology platforms often fail to fully consider the personalized needs of users at different time periods, resulting in the recommended hotels not fully meeting the expectations of users, thereby reducing the accuracy of hotel recommendations.

[0085] In order to solve the above problems, the present application provides a hotel recommendation method and device based on user behavior analysis to solve the problems existing in the above related technical problems. In the embodiment of the present application, on the one hand, the server recommends hotels by obtaining the user's travel standards and historical behavior data. The historical behavior data is collected in real time through a preset listener and reflects the user's attention to hotel features at different time points. By analyzing these behavior data, the server can more deeply understand the user's personalized needs and preference changes, so that the recommended hotel not only meets the user's travel standards, but also meets their implicit personal preferences. This method of combining user behavior data improves the accuracy of hotel recommendations; on the other hand, based on user behavior data, the server fits factor change curves of multiple preset hotel screening factors. These curves reveal the changes in the user's attention to different hotel attributes within a preset period. These factor change curves provide rich input features for the personal preference prediction model, so that the model can capture the dynamic changes in user preferences, thereby enhancing the personalization of hotel recommendations. The following is a detailed description using an exemplary embodiment.

[0086] The following will be combined with the attached Figure 1 -Attached Figure 6 , a hotel recommendation method based on user behavior analysis provided by an embodiment of the present application is introduced in detail. The method can be implemented by a computer program and can be run on a hotel recommendation device based on user behavior analysis based on the von Neumann system. The computer program can be integrated into an application or run as an independent tool application.

[0087] See also Figure 1 , which is a flowchart of a hotel recommendation method based on user behavior analysis, applied to a server. Figure 1 As shown, the method of the embodiment of the present application may include the following steps:

[0088] S101, receiving a hotel reservation request sent by a client, the hotel reservation request is used to request a server to generate an optimal hotel list belonging to the user, and the hotel reservation request carries a destination and a target user identifier;

[0089] Among them, the client refers to the device or software interface used by the user, through which a request can be sent to the server. This is usually a mobile application, a web interface, or other type of user interface. A hotel reservation request is a request made by a user through a client, with the purpose of allowing the server to generate a hotel list based on the user's needs. The server refers to the backend server or server cluster, which is responsible for processing the client's request, performing the necessary calculations, and returning a response to the client. The optimal hotel list is the hotel options provided by the server that best meet the user's needs and preferences. The destination is the place where the user plans to go, which can be a city, region or specific address, and is one of the key information in a hotel reservation request. The target user identifier is the information used to uniquely identify a user, which can be a username, user ID or other unique identification code.

[0090] In some embodiments, a user uses a hotel booking application on a smartphone, enters a travel destination "Paris XXX" and travel dates, and then clicks a search button to submit a booking request. This request is sent to the server, which contains the user's unique identifier "U12345678" and the destination "Paris XXX".

[0091] S102, acquiring a target travel standard corresponding to a target user identifier from a pre-configured mapping relationship between a user identifier and a user travel standard;

[0092] Among them, travel standards refer to the standards or restrictions set by a company or organization for employees on business trips regarding accommodation, transportation and other expenses, such as the maximum nightly limit for accommodation, reimbursable meal expenses, etc.

[0093] In some embodiments, in a large enterprise, an employee "Zhang San" has a work number "ZS123", which is his user ID in the system. The company sets accommodation and dining standards for each employee when traveling on business, which are stored in the system and mapped with the employee's user ID. When "Zhang San" needs to go on a business trip and uses his work number to make a hotel reservation on the system, the system will obtain the travel standard corresponding to "ZS123" from the pre-configured mapping relationship, that is, an accommodation limit of 1,000 yuan per night. In this way, the system can ensure that the hotel recommended for "Zhang San" meets his travel standards.

[0094] It should be noted that the mapping relationship between the pre-configured user ID and the user travel standard is configured by the backend engineer and can be continuously updated.

[0095] For example Figure 2 As shown, when the sales manager Zhang San (user ID: ZS123) needs to book a hotel, the system can quickly find his travel standards through his unique user ID, that is, an accommodation budget of 1,000 yuan per night, a food and beverage budget of 300 yuan per day, an economy class transportation standard, and other instructions without special requirements.

[0096] S103, screening hotel data within a preset range around the destination and meeting the target travel standards, and generating a set of candidate hotels;

[0097] The preset range refers to a geographical range around the destination, which is used to filter nearby hotels. The preset range can be a fixed distance (such as within 5 kilometers) or time (such as within 30 minutes' drive). The alternative hotel set refers to a list of hotels that meet the requirements based on the destination and target travel standards. These hotels can be used as potential options for users to book.

[0098] In some embodiments, for example, a user plans to travel to Shanghai to attend an important business meeting and uses the company's travel booking system to arrange accommodation. The user enters the meeting address as the destination in the system and provides his or her employee number to determine the travel criteria. The system filters out nearby hotels based on the preset range (within 5 kilometers of the meeting location) and the target travel criteria (for example, the budget for accommodation per night does not exceed 1,200 yuan, and only hotels that provide breakfast are considered). Finally, the system generates a set of alternative hotels containing 5 hotels that meet all the conditions for the user to select and book. This process ensures that users can quickly find accommodation options that are both convenient and in line with the company's travel policy.

[0099] S104, obtaining all historical user behavior data corresponding to the target user identifier before the current moment from the pre-collected dynamic mapping relationship between the user identifier and the user behavior information, wherein the user behavior information in the dynamic mapping relationship is stored one by one in chronological order and obtained by real-time monitoring of the target user identifier by a preset monitor;

[0100] The dynamic mapping relationship refers to the association between the user ID and its behavior information, which can be dynamically updated to reflect the user's latest behavior. The preset listener is a monitoring tool preset in the system, which is used to track and record user behavior in real time. Historical user behavior data is a record of user behavior in the past, which can be used to analyze the user's personalized needs.

[0101] In some embodiments, the user is identified by the user registration email "user123@example.com". The platform uses a preset listener to track and record all user behaviors on the platform in real time, including searched keywords, browsed hotel details pages, clicked links, etc. These behavior information are stored in the database one by one in chronological order, and associated with the user identifier "user123@example.com" to form a dynamic mapping relationship. When it is necessary to obtain the historical user behavior data of "user123@example.com", the system can retrieve all relevant records from this dynamic mapping relationship to obtain the historical user behavior data.

[0102] For example Figure 3 As shown in the figure, a JSON string is used to store all the behavior data of the user "user123@example.com". Each behavior record includes the behavior type, description, time, and the identifier and attributes of the related hotel.

[0103] S105, based on the user behavior data, fitting a factor change curve graph of multiple preset hotel screening factors of the target user to which the target user identifier belongs within a preset period, wherein the preset hotel screening factors are used to describe basic attributes and conditions of hotel characteristics, and the factor change curve graph is used to characterize changes in the degree of attention paid by the target user to each hotel screening factor at different time points;

[0104] Among them, user behavior data includes search keywords, browsed hotel details pages, and clicked hotel pictures or descriptions.

[0105] In some embodiments of the present application, based on user behavior data, the specific process of fitting a factor change curve graph of multiple preset hotel screening factors of a target user to whom a target user identification belongs within a preset period includes: obtaining multiple preset hotel screening factors; determining a preset period; drawing a two-dimensional coordinate graph based on the preset period and the multiple preset hotel screening factors; mapping the searched keywords, browsed hotel details pages, clicked hotel pictures or descriptions to the two-dimensional coordinate graph, and obtaining a factor change curve graph of multiple preset hotel screening factors of the target user to whom the target user identification belongs within the preset period.

[0106] Among them, a two-dimensional coordinate graph is a graphical representation method, usually with a horizontal axis and a vertical axis, used to show the changes of data over time or other variables. A factor change curve graph refers to a graph drawn on a two-dimensional coordinate graph based on user behavior data, showing the changes of user preferences for different hotel screening factors over time.

[0107] In the embodiment of the present application, by analyzing user behavior data and drawing a factor change curve, the user's preference changes for different hotel screening factors within a preset period can be accurately captured, thereby providing more personalized and accurate hotel recommendations, enhancing user experience, and improving booking conversion rates. This method enables the recommendation system to dynamically adapt to changes in user needs, ensure the relevance and timeliness of recommendation results, and ultimately bring users a more satisfactory travel booking experience.

[0108] S106, inputting factor change curves of multiple preset hotel screening factors into a pre-trained personal preference prediction model, and outputting a preference keyword sequence corresponding to the target user, where the preference keyword sequence is used to represent the user's personalized demand for hotels;

[0109] It should be noted that the preset hotel filter factors are the basic attributes or conditions used by the system to describe the characteristics of hotels. They are objective criteria used to filter and classify hotels, such as "price range", "hotel star", "location convenience", "user rating", etc. The preset hotel filter factors are usually used in database queries and hotel searches to help users narrow down their choices based on specific needs. These factors are often quantifiable and can be directly extracted from the hotel's data. The preset preference keyword set is a word or phrase that represents the user's personal preferences and preferences. The preference keyword set includes more specific descriptions, such as "free Wi-Fi", "sea view room", "Michelin restaurant", etc. These descriptions can capture the user's personalized needs for the hotel in more detail. The preset preference keyword set may be updated as user behavior and preferences change, while the preset hotel filter factors are relatively stable and are usually determined by the fixed attributes of the hotel.

[0110] Among them, for example Figure 4As shown, the pre-trained personal preference prediction model includes a region segmentation module, a preference strength extraction module, a context information collection module, a coding fusion module, and a random forest regression network.

[0111] In an embodiment of the present application, when inputting a factor change curve diagram of multiple preset hotel screening factors into a pre-trained personal preference prediction model, the specific process of outputting a preference keyword sequence corresponding to the target user includes: a region segmentation module segments the region where the hotel screening factor that appears most frequently in the factor change curve diagram of multiple preset hotel screening factors is located as a focus area, and the focus area is used to characterize the hotel attributes preferred by the target user in different time periods; the preference strength extraction module extracts and quantifies the target user's preference strength for the hotel screening factor based on the focus area; the context information collection module collects context information related to the target user, and the context information is used to characterize the user's hotel booking background; the encoding fusion module encodes and fuses the preference strength and context information to obtain a fusion feature; the fusion feature and the preset preference keyword set are input into the random forest regression network, and the preference keyword sequence corresponding to the target user is output.

[0112] In some embodiments of the present application, the specific process of extracting and quantifying the target user's preference strength for hotel screening factors based on the focus area includes: screening out the target factors located in the focus area from the preset hotel screening factor set; collecting the hotel star rating, price, user evaluation and booking frequency of the target user in the focus area to obtain historical booking data; and determining the comprehensive score of the target factor based on the historical booking data as the target user's preference strength for the hotel screening factor.

[0113] In the embodiments of the present application, for example Figure 5 As shown, a pre-trained personal preference prediction model is generated according to the following steps, specifically including: collecting historical hotel reservation data of each historical user; fitting a historical factor change curve of multiple preset hotel screening factors of each historical user within a preset period according to the historical hotel reservation data of each historical user; creating a personal preference prediction model, the personal preference prediction model includes a region segmentation module, a preference strength extraction module, a context information collection module, a coding fusion module and a random forest regression network; using the region segmentation module and the preference strength extraction module to process the historical factor change curve in turn to obtain the historical preference strength corresponding to each historical factor change curve; using the context information collection module to collect historical context information related to each historical user; using the coding fusion module to fuse the historical preference strength and its historical context information corresponding to each historical factor change curve to obtain the historical fusion feature of each historical user; based on the historical fusion feature of each historical user, performing machine learning on the random forest regression network to obtain the pre-trained personal preference prediction model.

[0114] Among them, the random forest regression network is an ensemble learning method, which belongs to a supervised learning algorithm and is mainly used to deal with regression problems, that is, predicting continuous numerical data. Random forest regression improves the accuracy and robustness of predictions by constructing multiple decision trees and aggregating their prediction results.

[0115] In the embodiment of the present application, by collecting and analyzing the user's historical hotel reservation data, using an advanced personal preference prediction model, the model combines regional segmentation, preference strength extraction, context information collection and encoding fusion technology, and uses a random forest regression network for machine learning, it can accurately fit and predict the user's hotel selection preferences. This method can provide highly personalized hotel recommendations for each user, thereby significantly improving user satisfaction and the accuracy of the reservation system, and ultimately enhancing user experience and improving conversion rates.

[0116] In some embodiments of the present application, based on the historical fusion features of each historical user, a random forest regression network is subjected to machine learning to obtain a pre-trained personal preference prediction model. The specific process includes: initializing a preset preference keyword set; in response to a data labeling instruction, selecting a target preference keyword corresponding to the data labeling instruction from the preset preference keyword set to perform data labeling on the historical fusion features of each historical user to obtain multiple model training samples; inputting each model training sample into the random forest regression network and outputting a model loss value; and generating a pre-trained personal preference prediction model based on the model loss value.

[0117] In some embodiments of the present application, the specific process of generating a pre-trained personal preference prediction model according to the model loss value includes: generating a pre-trained personal preference prediction model when the model loss value reaches the minimum; or, when the model loss value does not reach the minimum, updating the network parameters of the random forest regression network, and continuing to execute the step of inputting each model training sample into the random forest regression network until the loss value reaches the minimum; wherein,

[0118] The loss function of the random forest regression network is:

[0119]

[0120] in, is the model loss value, Indicates the first samples, represents the classification loss, Represents the random forest regression network for the Sample predicted preferred keywords, Expressing the The sample annotated preferred keyword labels, is the weight to balance the loss.

[0121] S107, from the candidate hotel set, all hotels that meet the preferred keyword sequence are selected as the optimal hotel list, and recommended to the client for display.

[0122] In some embodiments of the present application, the specific process of screening all hotels that meet the preference keyword sequence from the set of alternative hotels as the optimal hotel list includes: obtaining hotel information and historical user evaluation information of each alternative hotel in the set of alternative hotels; judging whether each preference keyword in the preference keyword sequence exists in the hotel information and historical user evaluation information of each alternative hotel, and obtaining a judgment result; from the judgment result, counting the number of each preference keyword in the preference keyword sequence that exists in the hotel information and historical user evaluation information of each alternative hotel; and generating an optimal hotel list based on the number of each preference keyword that exists in the preference keyword sequence.

[0123] Specifically, based on the number of preference keywords in the preference keyword sequence, the specific process of generating the optimal hotel list includes: obtaining hotel information whose number of preference keywords in the preference keyword sequence is greater than a preset threshold and alternative hotels corresponding to historical user evaluation information; arranging the obtained alternative hotels in descending order according to the number of preference keywords in their preference keyword sequence to obtain the optimal hotel list.

[0124] Furthermore, the best hotel list recommended to the client for display is as follows: Figure 6 shown.

[0125] In an embodiment of the present application, on the one hand, the server makes hotel recommendations by obtaining the user's travel standards and historical behavior data. The historical behavior data is collected in real time through a preset listener and reflects the user's attention to hotel features at different time points. By analyzing these behavior data, the server can have a deeper understanding of the user's personalized needs and preference changes, so that the recommended hotels not only meet the user's travel standards, but also meet their implicit personal preferences. This method of combining user behavior data improves the accuracy of hotel recommendations; on the other hand, based on user behavior data, the server fits factor change curves of multiple preset hotel screening factors. These curves reveal the changes in the user's attention to different hotel attributes within a preset period. These factor change curves provide rich input features for the personal preference prediction model, so that the model can capture the dynamic changes in user preferences, thereby enhancing the personalization of hotel recommendations.

[0126] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0127] See also Figure 7 , which shows a schematic diagram of the structure of a hotel recommendation device based on user behavior analysis provided by an exemplary embodiment of the present application. The hotel recommendation device based on user behavior analysis can be implemented as all or part of an electronic device through software, hardware or a combination of both. The device 1 includes a request receiving module 10, a travel standard acquisition module 20, a hotel data screening module 30, a user behavior data acquisition module 40, a factor change curve fitting module 50, a preference keyword sequence output module 60, and an optimal hotel list screening module 70.

[0128] The request receiving module 10 is used to receive a hotel reservation request sent by a client, the hotel reservation request is used to request the server to generate an optimal hotel list belonging to the user, and the hotel reservation request carries a destination and a target user identifier;

[0129] The travel standard acquisition module 20 is used to acquire the target travel standard corresponding to the target user identifier from the mapping relationship between the pre-configured user identifier and the user travel standard;

[0130] The hotel data screening module 30 is used to screen hotel data within a preset range around the destination and meeting the target travel standards, and generate a set of candidate hotels;

[0131] The user behavior data acquisition module 40 is used to acquire all historical user behavior data corresponding to the target user identity before the current moment from the dynamic mapping relationship between the pre-collected user identity and the user behavior information. The user behavior information in the dynamic mapping relationship is stored one by one in chronological order and obtained by real-time monitoring of the target user identity by a preset monitor;

[0132] The factor change curve fitting module 50 is used to fit the factor change curve of multiple preset hotel screening factors of the target user to which the target user identifier belongs within a preset period based on the user behavior data. The preset hotel screening factors are used to describe the basic attributes and conditions of hotel characteristics. The factor change curve is used to characterize the change in the degree of attention paid by the target user to each hotel screening factor at different time points;

[0133] The preference keyword sequence output module 60 is used to input the factor change curve diagrams of multiple preset hotel screening factors into the pre-trained personal preference prediction model, and output the preference keyword sequence corresponding to the target user, and the preference keyword sequence is used to represent the user's personalized demand for hotels;

[0134] The optimal hotel list screening module 70 is used to screen all hotels that meet the preferred keyword sequence from the candidate hotel set as the optimal hotel list, and recommend them to the client for display.

[0135] It should be noted that the hotel recommendation device based on user behavior analysis provided in the above embodiment only uses the division of the above functional modules as an example when executing the hotel recommendation method based on user behavior analysis. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the hotel recommendation device based on user behavior analysis provided in the above embodiment and the hotel recommendation method based on user behavior analysis belong to the same concept, and the implementation process thereof is detailed in the method embodiment, which will not be repeated here.

[0136] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0137] In an embodiment of the present application, on the one hand, the server makes hotel recommendations by obtaining the user's travel standards and historical behavior data. The historical behavior data is collected in real time through a preset listener and reflects the user's attention to hotel features at different time points. By analyzing these behavior data, the server can have a deeper understanding of the user's personalized needs and preference changes, so that the recommended hotels not only meet the user's travel standards, but also meet their implicit personal preferences. This method of combining user behavior data improves the accuracy of hotel recommendations; on the other hand, based on user behavior data, the server fits factor change curves of multiple preset hotel screening factors. These curves reveal the changes in the user's attention to different hotel attributes within a preset period. These factor change curves provide rich input features for the personal preference prediction model, so that the model can capture the dynamic changes in user preferences, thereby enhancing the personalization of hotel recommendations.

[0138] The present application also provides a computer-readable medium having program instructions stored thereon, which, when executed by a processor, implements the hotel recommendation method based on user behavior analysis provided by the above-mentioned various method embodiments.

[0139] The present application also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the hotel recommendation method based on user behavior analysis of the above-mentioned various method embodiments.

[0140] See also Figure 8 , is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 8 As shown, the electronic device 1000 may include: at least one processor 1001 , at least one network interface 1004 , a user interface 1003 , a memory 1005 , and at least one communication bus 1002 .

[0141] The communication bus 1002 is used to realize the connection and communication between these components.

[0142] The user interface 1003 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 1003 may also include a standard wired interface and a wireless interface.

[0143] The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0144] Among them, the processor 1001 may include one or more processing cores. The processor 1001 uses various interfaces and lines to connect various parts of the entire electronic device 1000, and executes various functions and processes data of the electronic device 1000 by running or executing instructions, programs, code sets or instruction sets stored in the memory 1005, and calling data stored in the memory 1005. Optionally, the processor 1001 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 1001 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 1001, and it can be implemented separately through a chip.

[0145] Among them, the memory 1005 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 1005 includes a non-transitory computer-readable storage medium. The memory 1005 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 1005 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 1005 may also be optionally at least one storage system located away from the aforementioned processor 1001. As Figure 8 As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a hotel recommendation application based on user behavior analysis.

[0146] exist Figure 8 In the electronic device 1000 shown, the user interface 1003 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 1001 can be used to call the hotel recommendation application based on user behavior analysis stored in the memory 1005, and specifically perform the following operations:

[0147] Receive a hotel reservation request sent by a client, where the hotel reservation request is used to request the server to generate an optimal hotel list belonging to the user, and the hotel reservation request carries a destination and a target user identifier;

[0148] Obtaining the target travel standard corresponding to the target user identifier from the mapping relationship between the pre-configured user identifier and the user travel standard;

[0149] Filter hotel data within a preset range around the destination that meets the target travel standards to generate a set of candidate hotels;

[0150] From the pre-collected dynamic mapping relationship between user IDs and user behavior information, all historical user behavior data corresponding to the target user ID before the current moment are obtained. The user behavior information in the dynamic mapping relationship is stored one by one in chronological order and obtained from real-time monitoring of the target user ID by a preset monitor;

[0151] Based on the user behavior data, a factor change curve of multiple preset hotel screening factors of the target user to which the target user ID belongs within a preset period is fitted. The preset hotel screening factors are used to describe the basic attributes and conditions of hotel characteristics. The factor change curve is used to characterize the change in the degree of attention paid by the target user to each hotel screening factor at different time points;

[0152] Inputting factor change curves of multiple preset hotel screening factors into a pre-trained personal preference prediction model, outputting a preference keyword sequence corresponding to the target user, the preference keyword sequence is used to represent the user's personalized demand for hotels;

[0153] From the set of candidate hotels, select all hotels that meet the preferred keyword sequence as the optimal hotel list and recommend them to the client for display.

[0154] In one embodiment, when the processor 1001 inputs the factor change curves of the plurality of preset hotel screening factors into the pre-trained personal preference prediction model and outputs the preference keyword sequence corresponding to the target user, the processor 1001 specifically performs the following operations:

[0155] The region segmentation module segments the region where the hotel screening factor with the highest frequency appears in the factor change curve diagram of multiple preset hotel screening factors as the focus region, and the focus region is used to represent the hotel attributes preferred by the target user in different time periods;

[0156] The preference strength extraction module extracts and quantifies the target user's preference strength for hotel screening factors based on the focus area;

[0157] The context information collection module collects context information related to the target user, and the context information is used to characterize the user's hotel booking background;

[0158] The encoding fusion module encodes and fuses the preference strength and context information to obtain fusion features;

[0159] The fused features and the preset preference keyword set are input into the random forest regression network to output the preference keyword sequence corresponding to the target user.

[0160] In one embodiment, when the processor 1001 extracts and quantifies the target user's preference strength for the hotel screening factor based on the focus area, the processor 1001 specifically performs the following operations:

[0161] From the preset hotel screening factor set, select the target factor located in the area of ​​interest;

[0162] Collect hotel star ratings, prices, user reviews, and booking frequencies of target users in the area of ​​interest to obtain historical booking data;

[0163] Based on historical booking data, the comprehensive score of the target factor is determined as the target user's preference intensity for the hotel screening factor.

[0164] In one embodiment, when the processor 1001 generates a pre-trained personal preference prediction model, the processor 1001 specifically performs the following operations:

[0165] Collect historical hotel booking data for each historical user;

[0166] According to the historical hotel booking data of each historical user, a historical factor change curve chart of multiple preset hotel screening factors of each historical user within a preset period is fitted;

[0167] Create a personal preference prediction model, which includes a region segmentation module, a preference strength extraction module, a context information collection module, a coding fusion module, and a random forest regression network;

[0168] The regional segmentation module and the preference strength extraction module are used to process the historical factor change curve graph in turn to obtain the historical preference strength corresponding to each historical factor change curve graph;

[0169] A context information collection module is used to collect historical context information related to each historical user;

[0170] The encoding fusion module is used to fuse the historical preference strength and historical context information corresponding to each historical factor change curve to obtain the historical fusion features of each historical user;

[0171] Based on the historical fusion features of each historical user, the random forest regression network is machine learned to obtain a pre-trained personal preference prediction model.

[0172] In one embodiment, when the processor 1001 performs machine learning on the random forest regression network based on the historical fusion features of each historical user to obtain a pre-trained personal preference prediction model, the following operations are specifically performed:

[0173] Initialize the preset favorite keyword set;

[0174] In response to the data annotation instruction, a target preference keyword corresponding to the data annotation instruction is selected from a preset preference keyword set to perform data annotation on the historical fusion features of each historical user to obtain a plurality of model training samples;

[0175] Input each model training sample into the random forest regression network and output the model loss value;

[0176] Based on the model loss value, a pre-trained personal preference prediction model is generated.

[0177] In one embodiment, when the processor 1001 generates a pre-trained personal preference prediction model according to the model loss value, the processor 1001 specifically performs the following operations:

[0178] When the model loss value reaches the minimum, a pre-trained personal preference prediction model is generated;

[0179] or,

[0180] When the model loss value has not reached the minimum, the network parameters of the random forest regression network are updated, and the step of inputting each model training sample into the random forest regression network is continued until the loss value reaches the minimum; wherein,

[0181] The loss function of the random forest regression network is:

[0182]

[0183] in, is the model loss value, Indicates the first samples, represents the classification loss, Represents the random forest regression network for the Sample predicted preferred keywords, Expressing the The sample annotated preferred keyword labels, is the weight to balance the loss.

[0184] In one embodiment, when the processor 1001 performs fitting of a factor change curve of a plurality of preset hotel screening factors of a target user to which the target user identifier belongs within a preset period based on the user behavior data, the following operations are specifically performed:

[0185] Get multiple preset hotel filter factors;

[0186] Determine the preset period;

[0187] Draw a two-dimensional coordinate graph based on a preset period and multiple preset hotel screening factors;

[0188] The searched keywords, the browsed hotel details pages, and the clicked hotel pictures or descriptions are mapped to a two-dimensional coordinate graph to obtain a factor change curve graph of multiple preset hotel screening factors of the target user to which the target user identification belongs within a preset period.

[0189] In one embodiment, when the processor 1001 performs the following operations when screening all hotels that meet the preferred keyword sequence from the candidate hotel set as the optimal hotel list:

[0190] Obtain hotel information and historical user evaluation information for each candidate hotel in the candidate hotel set;

[0191] Determine whether each preference keyword in the preference keyword sequence exists in the hotel information and historical user evaluation information of each candidate hotel, and obtain a determination result;

[0192] From the determination results, count the number of each preference keyword in the preference keyword sequence in the hotel information of each candidate hotel and the historical user evaluation information;

[0193] Based on the number of each preference keyword in the preference keyword sequence, an optimal hotel list is generated.

[0194] In one embodiment, when the processor 1001 generates the optimal hotel list based on the number of each preference keyword in the preference keyword sequence, the processor 1001 specifically performs the following operations:

[0195] Acquire hotel information in which the number of each preference keyword in the preference keyword sequence is greater than a preset threshold and candidate hotels corresponding to historical user evaluation information;

[0196] The obtained candidate hotels are arranged in descending order according to the number of preference keywords in the preference keyword sequence to obtain an optimal hotel list.

[0197] In an embodiment of the present application, on the one hand, the server makes hotel recommendations by obtaining the user's travel standards and historical behavior data. The historical behavior data is collected in real time through a preset listener and reflects the user's attention to hotel features at different time points. By analyzing these behavior data, the server can have a deeper understanding of the user's personalized needs and preference changes, so that the recommended hotels not only meet the user's travel standards, but also meet their implicit personal preferences. This method of combining user behavior data improves the accuracy of hotel recommendations; on the other hand, based on user behavior data, the server fits factor change curves of multiple preset hotel screening factors. These curves reveal the changes in the user's attention to different hotel attributes within a preset period. These factor change curves provide rich input features for the personal preference prediction model, so that the model can capture the dynamic changes in user preferences, thereby enhancing the personalization of hotel recommendations.

[0198] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program for hotel recommendation based on user behavior analysis can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium of the program for hotel recommendation based on user behavior analysis can be a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0199] The above disclosure is only the preferred embodiment of the present application, which certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A hotel recommendation method based on user behavior analysis, characterized in that: Applied to the server, the method includes: Receiving a hotel reservation request sent by a client, wherein the hotel reservation request is used to request the server to generate an optimal hotel list belonging to the user, and the hotel reservation request carries a destination and a target user identifier; Acquire the target travel standard corresponding to the target user identifier from the mapping relationship between the pre-configured user identifier and the user travel standard; Filter hotel data within a preset range around the destination and meeting the target travel criteria to generate a set of candidate hotels; Obtain all historical user behavior data corresponding to the target user identity before the current moment from the pre-collected dynamic mapping relationship between the user identity and the user behavior information, wherein the user behavior information in the dynamic mapping relationship is stored one by one in chronological order and obtained from real-time monitoring of the target user identity by a preset monitor; Based on the user behavior data, fitting a factor change curve graph of a plurality of preset hotel screening factors of the target user to which the target user identifier belongs within a preset period, wherein the preset hotel screening factors are used to describe basic attributes and conditions of hotel characteristics, and the factor change curve graph is used to characterize the change in the degree of attention paid by the target user to each hotel screening factor at different time points; Input the factor change curves of the plurality of preset hotel screening factors into a pre-trained personal preference prediction model, and output a preference keyword sequence corresponding to the target user, wherein the preference keyword sequence is used to characterize the user's personalized demand for hotels; wherein The pre-trained personal preference prediction model includes a region segmentation module, a preference strength extraction module, a context information collection module, a coding fusion module and a random forest regression network; wherein, The step of inputting the factor change curves of the plurality of preset hotel screening factors into a pre-trained personal preference prediction model and outputting a preference keyword sequence corresponding to the target user comprises: The region segmentation module segments the region where the hotel screening factor with the highest frequency appears in the factor change curve diagram of the plurality of preset hotel screening factors is located as a focus region, wherein the focus region is used to characterize the hotel attributes preferred by the target user in different time periods; The preference strength extraction module extracts and quantifies the preference strength of the target user for the hotel screening factor based on the focus area; The context information collection module collects context information related to the target user, wherein the context information is used to characterize the hotel booking background of the user; The encoding fusion module performs encoding fusion on the preference strength and the context information to obtain a fusion feature; Input the fusion features and the preset preference keyword set into a random forest regression network, and output a preference keyword sequence corresponding to the target user; From the candidate hotel set, all hotels that meet the preferred keyword sequence are screened as an optimal hotel list, and are recommended to the client for display.

2. The method according to claim 1, characterized in that The step of extracting and quantifying the target user's preference intensity for hotel screening factors based on the focus area includes: Filtering out target factors located in the area of ​​interest from a preset hotel screening factor set; Collect hotel star ratings, prices, user reviews, and booking frequencies of the target user in the area of ​​interest to obtain historical booking data; According to the historical booking data, a comprehensive score of the target factor is determined as the preference strength of the target user for the hotel screening factor.

3. The method according to claim 1, characterized in that The following steps are used to generate a pre-trained personal preference prediction model, including: Collect historical hotel booking data for each historical user; Fitting a historical factor change curve graph of a plurality of preset hotel screening factors of each historical user within a preset period according to the historical hotel booking data of each historical user; Creating a personal preference prediction model, the personal preference prediction model comprising a region segmentation module, a preference strength extraction module, a context information collection module, a coding fusion module, and a random forest regression network; The region segmentation module and the preference strength extraction module are used to process the historical factor change curve graph in sequence to obtain the historical preference strength corresponding to each historical factor change curve graph; Using the context information collection module to collect historical context information related to each historical user; The encoding fusion module is used to fuse the historical preference strength corresponding to each historical factor change curve graph and the historical context information to obtain the historical fusion feature of each historical user; Based on the historical fusion features of each historical user, the random forest regression network is machine-learned to obtain a pre-trained personal preference prediction model.

4. The method according to claim 3, characterized in that The method of performing machine learning on the random forest regression network based on the historical fusion features of each historical user to obtain a pre-trained personal preference prediction model includes: Initialize the preset favorite keyword set; In response to the data annotation instruction, a target preference keyword corresponding to the data annotation instruction is selected from the preset preference keyword set to perform data annotation on the historical fusion features of each historical user to obtain a plurality of model training samples; Input each model training sample into the random forest regression network and output the model loss value; A pre-trained personal preference prediction model is generated according to the model loss value.

5. The method according to claim 4, characterized in that Generating a pre-trained personal preference prediction model according to the model loss value includes: When the model loss value reaches a minimum, generating a pre-trained personal preference prediction model; or, When the model loss value has not reached the minimum, the network parameters of the random forest regression network are updated, and the step of inputting each model training sample into the random forest regression network is continued until the loss value reaches the minimum; wherein, The loss function of the random forest regression network is: in, is the model loss value, Indicates the first samples, represents the classification loss, Represents the random forest regression network for the Sample predicted preferred keywords, Expressing the The sample annotated preference keyword labels, is the weight to balance the loss.

6. The method according to claim 1, characterized in that The user behavior data includes search keywords, browsed hotel detail pages, clicked hotel pictures or descriptions; The step of fitting a factor change curve diagram of a plurality of preset hotel screening factors of a target user to which the target user identifier belongs within a preset period based on the user behavior data includes: Get multiple preset hotel filter factors; Determine the preset period; Draw a two-dimensional coordinate graph according to the preset period and multiple preset hotel screening factors; The searched keywords, the browsed hotel details pages, the clicked hotel pictures or descriptions are mapped to the two-dimensional coordinate graph to obtain a factor change curve graph of multiple preset hotel screening factors of the target user to which the target user identifier belongs within a preset period.

7. The method according to claim 1, characterized in that The step of selecting all hotels that meet the preferred keyword sequence from the candidate hotel set as the optimal hotel list includes: Obtain hotel information and historical user evaluation information of each candidate hotel in the candidate hotel set; Determine whether each preference keyword in the preference keyword sequence exists in the hotel information and historical user evaluation information of each candidate hotel, and obtain a determination result; From the determination result, counting the number of each favorite keyword in the favorite keyword sequence in the hotel information of each candidate hotel and the historical user evaluation information; An optimal hotel list is generated based on the number of each preference keyword present in the preference keyword sequence.

8. The method according to claim 7, characterized in that The step of generating an optimal hotel list based on the number of each favorite keyword in the favorite keyword sequence includes: Acquire hotel information in which the number of each preference keyword in the preference keyword sequence is greater than a preset threshold and candidate hotels corresponding to historical user evaluation information; The obtained candidate hotels are arranged in descending order according to the number of the preferred keywords in the preferred keyword sequence to obtain an optimal hotel list.

9. A hotel recommendation device based on user behavior analysis implemented by the method described in any one of claims 1 to 8, characterized in that: The device comprises: A request receiving module, used to receive a hotel reservation request sent by a client, wherein the hotel reservation request is used to request the server to generate an optimal hotel list belonging to the user, and the hotel reservation request carries a destination and a target user identifier; A travel standard acquisition module, used to acquire the target travel standard corresponding to the target user identifier from the mapping relationship between the pre-configured user identifier and the user travel standard; A hotel data screening module, used to screen hotel data within a preset range around the destination and meeting the target travel standards, and generate a set of candidate hotels; A user behavior data acquisition module is used to acquire all historical user behavior data corresponding to the target user identity before the current moment from the dynamic mapping relationship between the pre-collected user identity and the user behavior information, wherein the user behavior information in the dynamic mapping relationship is stored one by one in chronological order and obtained by real-time monitoring of the target user identity by a preset monitor; A factor change curve fitting module is used to fit a factor change curve of a plurality of preset hotel screening factors of a target user to which the target user identifier belongs within a preset period based on the user behavior data, wherein the preset hotel screening factors are used to describe basic attributes and conditions of hotel characteristics, and the factor change curve is used to characterize changes in the degree of attention paid by the target user to each hotel screening factor at different time points; A preference keyword sequence output module is used to input the factor change curve diagrams of the plurality of preset hotel screening factors into a pre-trained personal preference prediction model, and output a preference keyword sequence corresponding to the target user, wherein the preference keyword sequence is used to characterize the user's personalized demand for hotels; The optimal hotel list screening module is used to screen all hotels that meet the preferred keyword sequence from the candidate hotel set as the optimal hotel list, and recommend them to the client for display.

Images