Object recommendation method, medium and program product

Abstract

The embodiment of the invention relates to an object recommendation method, a medium and a program product. The method comprises the steps of obtaining target demand information corresponding to a target customer; obtaining respective target feature information of the plurality of target candidate objects; the target feature information comprises first-type feature information corresponding to specified-type features and second-type feature information corresponding to non-specified-type features; based on a text matching algorithm corresponding to the specified type of features and a vector matching algorithm corresponding to the non-specified type of features, performing matching processing on the target demand information and target feature information corresponding to the plurality of target candidate objects, and screening the target candidate objects to obtain a target recommendation object corresponding to the target client from the plurality of target candidate objects. According to the embodiment of the invention, the target recommendation object corresponding to the target customer can be efficiently and reliably screened out, the accuracy of the recommendation result is ensured, and the customer satisfaction is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to an object recommendation method, medium, and program product. Background Art

[0002] In the service industry, merchants or brokers often recommend desired objects to clients. Whether the recommendations meet their needs directly impacts customer satisfaction. However, the inventors have discovered that manual screening and recommendation methods are inefficient, while some technologies that rely on models to recommend objects often fail to meet customer needs. Therefore, a new object recommendation technology is urgently needed. Summary of the Invention

[0003] In order to solve the above technical problems or at least partially solve the above technical problems, the present disclosure provides an object recommendation method, medium and program product.

[0004] An embodiment of the present disclosure provides an object recommendation method, which includes: obtaining target demand information corresponding to a target customer; obtaining target feature information of each of a plurality of target candidate objects; wherein the target feature information includes first-category feature information corresponding to a specified type feature and second-category feature information corresponding to a non-specified type feature; based on a text matching algorithm corresponding to the specified type feature and a vector matching algorithm corresponding to the non-specified type feature, matching the target demand information with the target feature information corresponding to each of the plurality of target candidate objects, so as to screen out a target recommendation object corresponding to the target customer from the plurality of target candidate objects.

[0005] Optionally, obtaining the target demand information corresponding to the target customer includes: obtaining the target demand label corresponding to the target customer based on the customer database, and characterizing the target demand information corresponding to the target customer through the target demand label; wherein, the target demand label includes a label obtained based on the conversation data between the target customer and the object recommender.

[0006] Optionally, the target demand information is matched with the target feature information corresponding to each of the multiple target candidate objects to filter out the target recommendation object corresponding to the target customer from the multiple target candidate objects, including: using the text matching algorithm to perform text matching based on the first category of demand information in the target demand information and the first category feature information of each of the multiple target candidate objects to determine the first candidate object from the multiple target candidate objects; wherein the first category of demand information is information corresponding to the first category feature information; using the vector matching algorithm to perform vector matching based on the second category of demand information in the target demand information and the second category feature information of each of the first candidate objects to determine the first recommendation object from the first candidate object; wherein the second category of demand information is information corresponding to the second category feature information; based on the first recommendation object, obtain the target recommendation object corresponding to the target customer.

[0007] Optionally, obtaining the target recommended object corresponding to the target customer based on the first recommended object includes: when the number of the first recommended objects meets a preset number condition, using the first recommended object as the target recommended object corresponding to the target customer; when the number of the first recommended objects does not meet the preset number condition, selecting a second recommended object from multiple second candidate objects, and using both the first recommended object and the second recommended object as the target recommended objects corresponding to the target customer; wherein, the second candidate object is an object among the multiple target candidate objects other than the first candidate object.

[0008] Optionally, selecting a second recommended object from multiple second candidate objects includes: using the vector matching algorithm to perform vector matching based on the second type of demand information in the target demand information and the second type of feature information of each of the multiple second candidate objects to obtain the second recommended object.

[0009] Optionally, there are multiple target recommendation objects, and the method further includes: for each dimension of the multiple target dimensions, obtaining the matching score between the target customer and the target recommendation object corresponding to the dimension; wherein the target dimension includes at least one dimension corresponding to the specified type feature and at least one dimension corresponding to the non-specified type feature; obtaining the weights corresponding to each of the multiple target dimensions; based on the weights and matching scores corresponding to each of the multiple target dimensions, determining the weighted scores corresponding to each of the multiple target recommendation objects; based on the weighted scores corresponding to each of the multiple target recommendation objects, determining the target recommendation order corresponding to the multiple target recommendation objects.

[0010] Optionally, determining the target recommendation order corresponding to the multiple target recommendation objects based on the weighted scores corresponding to each of the multiple target recommendation objects includes: determining a first recommendation order corresponding to the multiple target recommendation objects based on the weighted scores corresponding to each of the multiple target recommendation objects; obtaining the object preference information of the target customer; and determining the target recommendation order of the multiple target recommendation objects based on the first recommendation order and the object preference information.

[0011] Optionally, obtaining the object preference information of the target customer includes: obtaining target behavior data corresponding to the target customer; and predicting the object preference information of the target customer based on the target behavior data.

[0012] Optionally, the first type of feature information is represented by text, and the second type of feature information is represented by a vector.

[0013] An embodiment of the present disclosure further provides a computer-readable storage medium, wherein the storage medium stores a computer program, and the computer program is used to execute the object recommendation method provided by the embodiment of the present disclosure.

[0014] The embodiments of the present disclosure further provide a computer program product, including a computer program, which, when executed by a processor, implements the object recommendation method provided in the embodiments of the present disclosure.

[0015] The above technical solution provided by the embodiment of the present disclosure can obtain the target demand information corresponding to the target customer, and obtain the target feature information of each of the multiple target candidate objects. The target feature information not only includes the first type of feature information corresponding to the specified type feature, but also further includes the second type of feature information corresponding to the non-specified type feature. Then, based on the text matching algorithm corresponding to the specified type feature and the vector matching algorithm corresponding to the non-specified type feature, the target demand information and the target feature information corresponding to each of the multiple target candidate objects can be matched, thereby screening the target recommendation object corresponding to the target customer from the multiple target candidate objects. The above method can further expand the non-specified type feature on the basis of the specified type feature of the target candidate object, so that the feature description of the target candidate object is more comprehensive and rich. On this basis, suitable matching algorithms will be adopted for different types of features, thereby efficiently and reliably screening the target recommendation object corresponding to the target customer from the multiple target candidate objects, ensuring the accuracy of the recommendation results, and helping to improve customer satisfaction.

[0016] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0018] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0019] Figure 1 A flowchart of an object recommendation method provided by an embodiment of the present disclosure;

[0020] Figure 2 A schematic diagram of a housing recommendation process provided by an embodiment of the present disclosure;

[0021] Figure 3 A flowchart of a method for determining a recommendation order provided in an embodiment of the present disclosure;

[0022] Figure 4 A schematic structural diagram of an object recommendation device provided by an embodiment of the present disclosure;

[0023] Figure 5 A schematic structural diagram of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features therein can be combined with each other in the absence of conflict.

[0025] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0026] After research, the inventors found that there are many problems with the relevant object recommendation technology. The following is an example of house recommendation: the existing house recommendation method considers fewer feature dimensions, and only considers basic features such as price and room type. Non-standard features such as the ventilation and lighting conditions of the house are basically not considered when recommending. These features require customers to actively communicate or view the house before they can be known and judged whether they meet the needs. Therefore, the existing recommendation results are difficult to meet customer needs more comprehensively. In addition, although some technologies can obtain customer needs, they mainly rely on static descriptions or questionnaires with customers, which are difficult to cope with changes in customer needs, resulting in the final recommendation results not meeting the customer's latest needs. In addition, related technologies also have problems such as insufficient personalized recommendation effects and insufficient privacy data protection. In order to improve at least one of the above problems, the embodiments of the present disclosure provide an object recommendation method, medium and program product, which are explained in detail below.

[0027] Figure 1 This is a flow chart of an object recommendation method provided by an embodiment of the present disclosure. The method can be executed by an object recommendation device, wherein the device can be implemented using software and / or hardware and can generally be integrated into an electronic device. Figure 1 As shown, the method mainly includes the following steps S102 to S106:

[0028] Step S102: obtaining target demand information corresponding to the target customer.

[0029] The target demand information is the target customer's demand information for the object to be recommended. The objects mentioned in the embodiments of the present disclosure include but are not limited to commodities such as houses and vehicles, and may also include tourist attractions, courses, etc. Any object that the customer needs to recommend is acceptable and is not restricted here.

[0030] The disclosed embodiments do not restrict the demand information of the objects required by the customers. The target demand information can be represented by multi-dimensional demands. Taking the demand for housing as an example, it includes but is not limited to basic demands such as house type, price range, geographical location, and area size. It can also include additional demands such as surrounding facilities, traffic conditions, ventilation conditions, lighting conditions, and suitability for leasing / investment / self-residence.

[0031] The embodiments of the present disclosure do not limit the specific method of obtaining target demand information. For example, the target demand information of the target customer can be retrieved from a database that pre-records customer needs, or the target demand information of the target customer can be determined based on the communication record between the target customer and the object recommender, and no limitation is imposed here. The above-mentioned object recommender can be a merchant, broker, planner, etc. who recommends the required object to the customer, and no limitation is imposed here.

[0032] Step S104 , obtaining target feature information of each of the plurality of target candidate objects; wherein the target feature information includes first-category feature information corresponding to the specified type feature and second-category feature information corresponding to the non-specified type feature.

[0033] In practical applications, designated type features can be determined based on object characteristics. These can be referred to as standard features or basic features. For example, in the case of a property, these can include basic parameter features such as price, location, room type, area, whether it's new, and floor level. For example, in the case of a vehicle, these can include features such as price, brand, and vehicle model. These designated type features are standard and objective and can be accurately represented using short text or other characters. Non-designated type features, which are features other than designated type features, can also be referred to as non-standard features or extended features. For example, in the case of a property, features such as ventilation, noise levels, lighting conditions, surrounding environment, and suitability for investment or leasing are all non-designated type features. Understandably, these features are often difficult to accurately represent using short text or other characters. They may require longer text to describe, and may be subjective or require a degree of measurement, making them difficult to represent using simple absolute terms such as "yes / no" or "yes / no." Compared with the related art that only considers specified type features, the embodiment of the present disclosure not only considers specified type features, but also non-specified type features, which more comprehensively expands the description dimension of the object so that objects that meet customer needs in all aspects can be found more accurately in the future.

[0034] Taking into account that specified type features can usually be accurately expressed through short texts, while non-specified type features may require longer texts to describe, in order to be able to process the above features more accurately in the future, in some embodiments, the first type of feature information corresponding to the specified type features is represented by text, and the second type of feature information corresponding to the non-specified type features is represented by vectors. The vector method can use high-dimensional information to more accurately and reliably present the meaning of non-specified type features, and can better present deep semantic information, which is convenient for subsequent matching.

[0035] In actual applications, the above-mentioned target feature information can be obtained in advance through multiple sources and entered into the object database. For example, the target feature information can be obtained through multiple channels such as visiting the object seller, the actual evaluation object, and a third-party data source. The object feature information stored in the object database can be made as rich and comprehensive as possible. When it is necessary to recommend an object to the target customer, it can be directly retrieved from the object database. The rich and comprehensive object feature information is more helpful in meeting customer needs.

[0036] Step S106, based on the text matching algorithm corresponding to the specified type features and the vector matching algorithm corresponding to the non-specified type features, the target demand information is matched with the target feature information corresponding to each of the multiple target candidate objects to filter out the target recommendation object corresponding to the target customer from the multiple target candidate objects.

[0037] In practical applications, the first type of demand information corresponding to the specified type of features can be first extracted from the target demand information, and the second type of demand information corresponding to the non-specified type of features can be extracted from the target demand information. Then, in some examples, the first type of demand information can be used to perform text matching with the first type of feature information corresponding to each of the multiple target candidate objects, and the second type of demand information can be used to perform vector matching with the second type of feature information corresponding to each of the multiple target candidate objects to obtain target matching scores corresponding to each of the multiple target candidate objects. The target matching scores can be obtained by weighting the text matching scores and the vector matching scores. Then, based on the target matching scores corresponding to each of the multiple target candidate objects, the target recommended objects corresponding to the target customers are screened from the multiple target candidate objects. In other examples, considering the accuracy of the first type of feature information, the first type of demand information can be used to perform text matching with the first type of feature information corresponding to each of the multiple target candidate objects, and some candidate objects can be filtered out first. Then, the second type of demand information can be used to perform vector matching with the second type of feature information corresponding to each of the remaining candidate objects to obtain target recommended objects.

[0038] For ease of understanding, let's take the candidate object as an example, see Figure 2 The figure shows a process flow diagram for property recommendation, illustrating how a client can initiate a property recommendation request to a server, which then returns the target recommended property. The figure further illustrates the server's object recommendation mechanism. Specifically, the server can obtain customer demand information and property feature information, perform text matching on the first category of demand information (also called standard demand information) in the customer demand information with the first category of feature information (also called standard feature information) in the property feature information, and perform vector matching on the second category of demand information (also called non-standard demand information) in the customer demand information with the second category of feature information (also called non-standard feature information) in the property feature information, to obtain the target recommended property.

[0039] The disclosed embodiment can further expand the non-specified type features on the basis of the specified type features of the target candidate objects through the above-mentioned method, so that the feature description of the target candidate objects is more comprehensive and rich. On this basis, it will also adopt respective suitable matching algorithms for different types of features, so as to efficiently and reliably screen out the target recommendation objects corresponding to the target customers from multiple target candidate objects, ensure the accuracy of the recommendation results, and help improve customer satisfaction.

[0040] In some embodiments, step S106, i.e., obtaining target demand information corresponding to the target customer, can be performed by referring to the following steps: obtaining a target demand tag corresponding to the target customer based on the customer database, and representing the target demand information corresponding to the target customer using the target demand tag; wherein the target demand tag includes a tag obtained based on the conversation data between the target customer and the object recommender. A detailed explanation is provided below:

[0041] In actual applications, the customer database can pre-store demand tags corresponding to each customer and can also regularly update the demand tags. If the object is a house, the demand tags can include: a tag indicating the house type, a tag indicating the floor of the house, a tag indicating the house's geographical location, a tag indicating the house's orientation, a tag indicating the house's price, etc. In this way, it is easier to retrieve the target customer's needs in a timely manner when recommending objects to the target customer, and the format of the demand tags is more convenient for rapid analysis and processing, so as to promptly recommend objects that meet the customer's needs. The target demand tags include tags obtained based on the conversation data between the target customer and the object recommender. The number of target demand tags can be one or more, and there is no limit here. The target customer and the recommender can communicate via text through a specific application or webpage, thereby obtaining a textual communication record. Alternatively, the target customer and the recommender can communicate via voice through a specific application, webpage, or phone, converting the audio into a textual communication record. Based on the communication record, conversation data can be obtained. This conversation data can include the latest communication record between the target customer and the recommender (i.e., the most recent communication record since the current moment), as well as historical communication records between the target customer and the recommender, such as the N communication records prior to the latest communication record between the target customer and the recommender. Based on the most recent communication record, the target customer's current demand label can be obtained, and based on historical communication records, the target customer's historical demand labels can be obtained. Both the current demand label and the historical demand label can be used as the target demand label, or only the current demand label can be used as the target demand label, with the specific setting being flexible. In specific implementations, real-time feedback mechanisms such as natural language understanding (NLP) processing and large models can be introduced. The model can be dynamically adjusted based on the customer's immediate responses and interactive information in different scenarios, accurately analyzing conversation data, better identifying customer intent, and generating labels that best meet customer needs, thereby improving the accuracy and timeliness of customer needs.

[0042] By converting conversation data into structured demand tags, we can present user needs in a concise and clear manner, making it easier to quickly retrieve matching objects. Furthermore, this approach ensures that target demand information fully reflects the customer's current needs. For example, the latest communication records can reflect the customer's latest needs in real time. Even if customer needs change, we can better respond and make real-time object recommendations based on these changes, helping to provide the objects that best meet the target customer's current needs.

[0043] To protect customer privacy, customer data stored in the customer database is processed using a specific encryption algorithm. This disclosure does not restrict the specific encryption algorithm used. Furthermore, privacy-preserving technologies such as federated learning can also be used during the processing of customer data. This approach effectively improves data security and prevents potential privacy leaks.

[0044] In some embodiments, step S106, i.e., matching the target demand information with the target feature information corresponding to each of the plurality of target candidate objects to screen the target recommendation objects corresponding to the target customer from the plurality of target candidate objects, may be performed with reference to steps A to C below:

[0045] Step A utilizes a text matching algorithm to perform text matching based on the first-category requirement information in the target requirement information and the first-category feature information of each of the multiple target candidate objects, thereby determining a first candidate object from the multiple target candidate objects. The first-category requirement information is information corresponding to the first-category feature information. For example, the first-category requirement information may include price requirements, location requirements, room type requirements, area requirements, new / second-hand housing requirements, floor requirements, etc. Since both the first-category requirement information and the first-category feature information can be represented by brief text or characters, an accurate matching result can be obtained by directly utilizing a text matching algorithm. In practical applications, the above-described text matching algorithm can be used to calculate a first matching score between the first-category requirement information in the target requirement information and the first-category feature information of each of the multiple target candidate objects. If the first-category feature information includes information of multiple different specified feature types, the first matching score can be obtained by weighting the matching scores corresponding to the various specified feature types. Subsequently, the first candidate object can be determined from the multiple target candidate objects based on the first matching scores of each of the multiple target candidate objects. For example, a target candidate object whose first matching score exceeds a preset first score threshold can be selected as the first candidate object.

[0046] It is understandable that the above method of performing text matching for specified type features is not only simple but also highly accurate. Through the above method, the first candidate objects whose specific type features meet customer needs can be preferentially screened out.

[0047] Step B, using a vector matching algorithm, performs vector matching based on the second category demand information in the target demand information and the second category feature information of each of the first candidate objects to determine the first recommended object from the first candidate objects; wherein the second category demand information is information corresponding to the second category feature information.

[0048] In the case of screening out a first candidate object whose specific type of features meet the customer's needs, vector matching is performed on non-specified type features, and semantic matching can be achieved at a deeper level with the help of high-dimensional features, so that the customer's second type of demand information and the second type of feature information of each of the first candidate objects can be reasonably and effectively matched. For example, through the above-mentioned vector matching algorithm, the second matching score between the second type of demand information in the target demand information and the second type of feature information of each of the first candidate objects can be calculated. If the second type of feature information contains information of multiple different non-specified type features, the second matching score can be obtained by weighting the matching scores corresponding to the various non-specified type features. Afterwards, the first recommended object can be determined from the first candidate objects based on the high or low second matching scores of each of the first candidate objects. For example, the first candidate object whose second matching score is higher than the preset second score threshold can be used as the first recommended object.

[0049] Step C: Based on the first recommended object, obtain the target recommended object corresponding to the target customer. In actual applications, the first recommended object can be directly used as the target recommended object. Alternatively, if the number of first recommended objects does not meet the requirements, other methods can be used to determine additional recommended objects to ensure that the number of target recommended objects meets the quantity constraint, thereby providing the target customer with more object options and ensuring a better customer experience.

[0050] In some specific implementation examples, step C can be performed with reference to the following steps C1 and C2:

[0051] In step C1, if the number of first recommended objects meets a preset quantity condition, the first recommended object is used as a target recommended object for the target customer. The present embodiment does not limit the preset quantity condition. For example, the preset quantity condition may indicate that the number of objects recommended for the target customer must reach a target quantity. The specific value of the target quantity can be flexibly set according to needs.

[0052] Step C2: When the number of first recommended objects does not meet the preset number condition, select a second recommended object from multiple second candidate objects, and use both the first recommended object and the second recommended object as target recommended objects corresponding to the target customer; wherein the second candidate object is an object other than the first candidate object among the multiple target candidate objects.

[0053] That is, if the number of first recommended objects is small, in order to provide the target customer with more options and ensure user experience, the embodiment of the present disclosure can also obtain a second recommended object from multiple second candidate objects, that is, try to select a second recommended object from the multiple target candidate objects other than the first candidate object. The embodiment of the present disclosure does not limit the method of obtaining the additional second recommended object. The method of screening the second recommended object from the multiple second candidate objects can be the same as or different from the method of screening the first recommended object, and is not limited here. In some specific implementation examples, the step of selecting the second recommended object from the multiple second candidate objects in the above step C2 includes: using a vector matching algorithm to perform vector matching based on the second type of demand information in the target demand information and the second type of feature information of each of the multiple second candidate objects to obtain the second recommended object. It can be understood that the aforementioned first candidate object is the object obtained by text matching screening. If the number of recommended objects obtained as a whole is insufficient, you can try to use a vector matching method to screen the second recommended object that may meet the user's non-standard needs from the multiple target candidate objects other than the first candidate object, and then recommend the first recommended object and the second recommended object together to the target customer for selection.

[0054] In the case where there are multiple target recommendation objects, the object recommendation method provided by the embodiment of the present disclosure further includes the following steps a to d:

[0055] Step a: For each of the multiple target dimensions, obtain the matching score between the target customer and the target recommendation object corresponding to that dimension. The target dimensions include at least one dimension corresponding to a specified type of feature and at least one dimension corresponding to a non-specified type of feature. The target dimensions can be flexibly set based on needs and are not limited here.

[0056] Step b: Obtain the weights corresponding to the multiple target dimensions. The weights corresponding to the target dimensions can be flexibly set according to needs, and the weights can also be adjusted dynamically.

[0057] Step c: Determine the weighted scores corresponding to the multiple target recommendation objects based on the weights and matching scores corresponding to the multiple target dimensions. It is understandable that the target recommendation objects obtained in the aforementioned manner may be objects obtained successively through the text matching method corresponding to the specified type features and the vector matching method corresponding to the non-specified type features. In order to reasonably display the order of the target recommendation objects to the customer, the embodiment of the present disclosure may perform weighted processing based on the dimensions corresponding to the specified type features and the dimensions corresponding to the non-specified type features, and comprehensively present the performance of the target recommendation objects in the above-mentioned dimensions through the weighted scores of the target recommendation objects.

[0058] Step d, based on the weighted scores corresponding to the multiple target recommendation objects, determine the target recommendation order corresponding to the multiple target recommendation objects. Specifically, the target recommendation order corresponding to the multiple target recommendation objects can be obtained based on the order of arranging the weighted scores from high to low. Afterwards, the object information of the multiple target recommendation objects can be presented on the client interface according to the target recommendation order. It can be understood that the higher the weighted score, the greater the match with the target customer's needs. Recommending objects to target customers in the above manner helps target customers to view the object information that best meets their needs in the first place, which not only saves customers' time and energy, but also improves customer satisfaction.

[0059] In some specific implementation examples, step d may be performed with reference to the following steps d1 to d3:

[0060] Step d1: Determine a first recommendation order for the target recommendation objects based on their respective weighted scores. In some embodiments, the first recommendation order is the order in which the target recommendation objects are arranged in descending order based on their respective weighted scores.

[0061] Step d2: Obtain the target customer's object preference information. The object preference information can be preference information provided by the customer or preference information obtained based on customer data analysis such as customer behavior data. In some specific implementation examples, step d2 can be performed with reference to the following steps d2.1 and d2.2:

[0062] Step d2.1: Obtain target behavior data corresponding to the target customer. Target behavior data is a record of behavior associated with an object. Target behavior data includes, but is not limited to, online behavior records of the target customer, such as browsing and adding to a collection on the target application or target page. It may also include offline behavior records, such as visitor behavior. For example, if the object is a property, the specific behavior information may include viewing behavior information. This includes, but is not limited to, online viewing behavior such as browsing and adding to a collection on the target application or target page. It may also include offline viewing behavior information, such as viewing trajectories.

[0063] Step d2.2: Predict the target customer's object preference information based on the target behavior data. The disclosed embodiments can determine object preference information based on the target behavior data. For example, if the common characteristics of the properties viewed and collected by the target customer are second-hand properties and properties with elevators, this indicates that the target customer prefers to purchase second-hand properties and requires elevators. Using this target behavior data, the target customer's object preference information can be reasonably and reliably determined.

[0064] Step d3: determining a target recommendation order of multiple target recommendation objects based on the first recommendation order and the object preference information.

[0065] The disclosed embodiments can adjust the first recommendation order based on object preference information. For example, if the first recommendation order is indicated from front to back as: Listing 1 to Listing 3, and these three properties are all second-hand properties, but Listing 1 and Listing 3 are walk-up properties, and Listing 2 is an elevator property, then the order can be adjusted to: Listing 2, Listing 1, Listing 3. This approach can fully incorporate customer preferences to achieve personalized recommendations, ensure the rationality of the order of objects recommended to customers, and prioritize objects that meet customer preferences.

[0066] For ease of understanding, in some specific implementation examples, you can refer to Figure 3 The flowchart of a method for determining a recommendation order shown mainly includes the following steps S302 to S314:

[0067] Step S302: For each of the multiple target dimensions, obtain a matching score between the target customer and the target recommendation object corresponding to the dimension; wherein the target dimension includes at least one dimension corresponding to a specified type feature and at least one dimension corresponding to a non-specified type feature;

[0068] Step S304: Obtain weights corresponding to the multiple target dimensions;

[0069] Step S306: Determine weighted scores corresponding to the multiple target recommendation objects based on the weights and matching scores corresponding to the multiple target dimensions;

[0070] Step S308: determining a first recommendation order corresponding to the plurality of target recommendation objects based on the weighted scores corresponding to the plurality of target recommendation objects;

[0071] Step S310, obtaining target behavior data corresponding to the target customer;

[0072] Step S312: predicting the target customer's object preference information based on the target behavior data.

[0073] Step S314: determining a target recommendation order of the plurality of target recommendation objects based on the first recommendation order and the object preference information.

[0074] Through the above method, on the basis of obtaining the target recommendation object, we can fully combine the performance of the target recommendation object in the specified type characteristics, non-specified type characteristics and the customer's preference information, and reasonably determine the target recommendation order of multiple target recommendation objects, so as to achieve personalized recommendations that truly meet customer needs and improve the efficiency of object recommendation.

[0075] In summary, the above-mentioned object recommendation method provided by the embodiment of the present disclosure can further expand the non-specified type features on the basis of the specified type features of the target candidate objects, so that the feature description of the target candidate objects is more comprehensive and rich. On this basis, it will also adopt respective suitable matching algorithms for different types of features, so as to efficiently and reliably filter out the target recommended objects corresponding to the target customers from multiple target candidate objects, and ensure the accuracy of the recommendation results. Furthermore, it can also accurately know the real-time needs of customers based on session data, have the ability to better respond to changes in customer needs, and represent customer needs through structured tags, which is more convenient to process, so as to timely recommend the objects that best meet the current needs to customers. Furthermore, combined with customer preference information, personalized recommendations can be provided to customers more accurately, ensuring the rationality of the ranking of multiple objects recommended to customers, promoting customers to select objects that meet their needs as soon as possible, and comprehensively improving customer satisfaction. In addition, the embodiment of the present disclosure can use privacy technologies such as encryption algorithms for data processing, which can better ensure data security and protect customer privacy.

[0076] Corresponding to the aforementioned object recommendation method, the present disclosure further provides an object recommendation device, Figure 4 This is a schematic diagram of the structure of an object recommendation device provided by an embodiment of the present disclosure. The device can be implemented by software and / or hardware and can generally be integrated into an electronic device, such as Figure 4 As shown, the object recommendation device includes:

[0077] The customer demand acquisition module 402 is used to obtain target demand information corresponding to target customers;

[0078] The object feature acquisition module 404 is configured to acquire target feature information of each of the plurality of target candidate objects; wherein the target feature information includes first-category feature information corresponding to the specified type feature and second-category feature information corresponding to the non-specified type feature;

[0079] The matching recommendation module 406 is used to match the target demand information with the target feature information corresponding to each of the multiple target candidate objects based on a text matching algorithm corresponding to the specified type feature and a vector matching algorithm corresponding to the non-specified type feature, so as to screen out the target recommended object corresponding to the target customer from the multiple target candidate objects.

[0080] The above-mentioned device can further expand the non-specified type features based on the specified type features of the target candidate objects, making the feature description of the target candidate objects more comprehensive and rich. On this basis, it will also adopt suitable matching algorithms for different types of features, so as to efficiently and reliably screen out the target recommendation objects corresponding to the target customers from multiple target candidate objects, ensure the accuracy of the recommendation results, and help improve customer satisfaction.

[0081] In some embodiments, the customer demand acquisition module 402 is specifically used to: obtain the target demand label corresponding to the target customer based on the customer database, and represent the target demand information corresponding to the target customer through the target demand label; wherein, the target demand label includes a label obtained based on the conversation data between the target customer and the object recommender.

[0082] In some implementations, the customer data stored in the customer database is data processed by a specific encryption algorithm.

[0083] In some embodiments, the matching recommendation module 406 is specifically used to: use the text matching algorithm to perform text matching based on the first category of demand information in the target demand information and the first category of feature information of each of the multiple target candidate objects, so as to determine the first candidate object from the multiple target candidate objects; wherein the first category of demand information is information corresponding to the first category of feature information; use the vector matching algorithm to perform vector matching based on the second category of demand information in the target demand information and the second category of feature information of each of the first candidate objects, so as to determine the first recommended object from the first candidate object; wherein the second category of demand information is information corresponding to the second category of feature information; based on the first recommended object, obtain the target recommended object corresponding to the target customer.

[0084] In some embodiments, the matching recommendation module 406 is specifically used to: when the number of the first recommended objects meets a preset number condition, use the first recommended object as the target recommended object corresponding to the target customer; when the number of the first recommended objects does not meet the preset number condition, select a second recommended object from multiple second candidate objects, and use both the first recommended object and the second recommended object as the target recommended objects corresponding to the target customer; wherein, the second candidate object is an object among the multiple target candidate objects other than the first candidate object.

[0085] In some embodiments, the matching recommendation module 406 is specifically configured to: utilize the vector matching algorithm to perform vector matching based on the second type of demand information in the target demand information and the second type of feature information of each of the plurality of second candidate objects to obtain a second recommended object.

[0086] In some embodiments, there are multiple target recommendation objects, and the device also includes a recommendation order determination module, which is used to: for each dimension of the multiple target dimensions, obtain the matching score between the target customer and the target recommendation object corresponding to the dimension; wherein the target dimension includes at least one dimension corresponding to the specified type feature and at least one dimension corresponding to the non-specified type feature; obtain the weights corresponding to each of the multiple target dimensions; based on the weights and matching scores corresponding to each of the multiple target dimensions, determine the weighted scores corresponding to each of the multiple target recommendation objects; based on the weighted scores corresponding to each of the multiple target recommendation objects, determine the target recommendation order corresponding to the multiple target recommendation objects.

[0087] In some embodiments, the order determination module is specifically used to: determine a first recommendation order corresponding to the multiple target recommendation objects based on the weighted scores corresponding to each of the multiple target recommendation objects; obtain the object preference information of the target customer; and determine a target recommendation order of the multiple target recommendation objects based on the first recommendation order and the object preference information.

[0088] In some implementations, the sequence determination module is specifically configured to: obtain target behavior data corresponding to the target customer; and predict the object preference information of the target customer based on the target behavior data.

[0089] In some embodiments, the first type of feature information is represented by text, and the second type of feature information is represented by a vector.

[0090] The object recommendation device provided in the embodiments of the present disclosure can execute the object recommendation method provided in any embodiment of the present disclosure, and has the corresponding functional modules and beneficial effects of the execution method.

[0091] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described device embodiment can refer to the corresponding process in the method embodiment, and will not be repeated here.

[0092] An embodiment of the present disclosure provides an electronic device, which includes: a storage device storing a computer program; and a processing device configured to execute the computer program in the storage device to implement the steps of any one of the methods in the present disclosure.

[0093] Reference below Figure 5 , which shows a schematic structural diagram of an electronic device 500 suitable for implementing the embodiments of the present disclosure. The terminal devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 5 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0094] like Figure 5 As shown, the electronic device 500 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 into a random access memory (RAM) 503. Various programs and data required for the operation of the electronic device 500 are also stored in the RAM 503. The processing device 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0095] Typically, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509. The communication device 509 may allow the electronic device 500 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 5 The electronic device 500 is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead.

[0096] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program carried on a non-transitory computer-readable medium, the computer program including program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device 509, or installed from a storage device 508, or installed from a ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the method of the embodiment of the present disclosure are performed.

[0097] In addition to the above-mentioned methods and devices, the embodiments of the present disclosure may also be a computer program product, which includes computer program instructions, which, when executed by a processor, cause the processor to perform the method provided by the embodiments of the present disclosure. The computer program product can be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present disclosure, and the programming languages ​​include object-oriented programming languages ​​such as Java, C++, etc., and also include conventional procedural programming languages ​​such as "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0098] In addition, the embodiment of the present disclosure may also be a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the processor is enabled to execute the object recommendation method provided by the embodiment of the present disclosure.

[0099] The computer-readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can, for example, include but is not limited to a system, device or component of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0100] The embodiments of the present disclosure further provide a computer program product, including a computer program / instruction, which implements the object recommendation method in the embodiments of the present disclosure when executed by a processor.

[0101] It is understandable that before using the technical solutions disclosed in the various embodiments of this disclosure, the type, scope of use, usage scenarios, etc. of the personal information involved in this disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0102] For example, in response to a user's active request, a prompt message is sent to the user to clearly inform the user that the operation requested will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the electronic device, application, server, storage medium, or other software or hardware that performs the operations of the disclosed technical solution based on the prompt message.

[0103] As an optional but non-limiting implementation, in response to receiving a user's active request, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. Furthermore, the pop-up window may also contain a selection control for the user to select "agree" or "disagree" to provide personal information to the electronic device.

[0104] It is understandable that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0105] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0106] The foregoing description is intended only to provide specific embodiments of the present disclosure, intended to enable those skilled in the art to understand and implement the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the embodiments described herein, but rather to be construed in the broadest manner consistent with the principles and novel features disclosed herein.

Claims

1. An object recommendation method, characterized in that: include: Obtain target demand information corresponding to target customers; Acquire target feature information of each of the plurality of target candidate objects; wherein the target feature information includes first-category feature information corresponding to the specified type feature and second-category feature information corresponding to the non-specified type feature; Based on the text matching algorithm corresponding to the specified type features and the vector matching algorithm corresponding to the non-specified type features, the target demand information is matched with the target feature information corresponding to each of the multiple target candidate objects to filter out the target recommendation object corresponding to the target customer from the multiple target candidate objects.

2. The method according to claim 1, characterized in that The obtaining of target demand information corresponding to the target customer includes: A target demand label corresponding to a target customer is obtained based on a customer database, and the target demand information corresponding to the target customer is represented by the target demand label; wherein the target demand label includes a label obtained based on conversation data between the target customer and an object recommender.

3. The method according to claim 1, characterized in that The matching process of the target demand information with the target feature information corresponding to each of the plurality of target candidate objects to screen out a target recommendation object corresponding to the target customer from the plurality of target candidate objects includes: Using the text matching algorithm, text matching is performed based on the first type of requirement information in the target requirement information and the first type of feature information of each of the multiple target candidate objects, so as to determine a first candidate object from the multiple target candidate objects; wherein the first type of requirement information is information corresponding to the first type of feature information; Using the vector matching algorithm, performing vector matching based on the second type of demand information in the target demand information and the second type of feature information of each of the first candidate objects, so as to determine a first recommended object from the first candidate objects; wherein the second type of demand information is information corresponding to the second type of feature information; Based on the first recommended object, a target recommended object corresponding to the target customer is obtained.

4. The method according to claim 3, characterized in that The obtaining, based on the first recommended object, a target recommended object corresponding to the target customer includes: In the case that the number of the first recommended objects meets a preset number condition, the first recommended objects are used as target recommended objects corresponding to the target customers; When the number of the first recommended objects does not meet the preset number condition, a second recommended object is selected from multiple second candidate objects, and both the first recommended object and the second recommended object are used as target recommended objects corresponding to the target customer; wherein, the second candidate object is an object among the multiple target candidate objects other than the first candidate object.

5. The method according to claim 4, characterized in that The selecting a second recommended object from a plurality of second candidate objects includes: The vector matching algorithm is used to perform vector matching based on the second type of demand information in the target demand information and the second type of feature information of each of the plurality of second candidate objects to obtain a second recommended object.

6. The method according to claim 1, wherein There are multiple target recommendation objects, and the method further includes: For each dimension of the plurality of target dimensions, obtaining a matching score between the target customer and the target recommendation object corresponding to the dimension; wherein the target dimension includes at least one dimension corresponding to the specified type feature and at least one dimension corresponding to the non-specified type feature; Obtaining weights corresponding to each of the multiple target dimensions; Determine the weighted scores corresponding to the multiple target recommendation objects based on the weights and matching scores corresponding to the multiple target dimensions; Based on the weighted scores corresponding to the multiple target recommendation objects, a target recommendation order corresponding to the multiple target recommendation objects is determined.

7. The method according to claim 6, characterized in that The determining the target recommendation order corresponding to the multiple target recommendation objects based on the weighted scores corresponding to the multiple target recommendation objects includes: Determining a first recommendation order corresponding to the multiple target recommendation objects based on the weighted scores corresponding to the multiple target recommendation objects; Obtaining object preference information of the target customer; Based on the first recommendation order and the object preference information, a target recommendation order of the plurality of target recommendation objects is determined.

8. The method according to claim 7, characterized in that The obtaining of the target customer's object preference information includes: Obtaining target behavior data corresponding to the target customer; The object preference information of the target customer is predicted based on the target behavior data.

9. The method according to any one of claims 1 to 8, characterized in that The first type of feature information is represented by text, and the second type of feature information is represented by a vector.

10. A computer-readable storage medium, characterized in that The storage medium stores a computer program, and the computer program is used to execute the object recommendation method according to any one of claims 1 to 9.

11. A computer program product, characterized in that The invention comprises a computer program, which implements the object recommendation method according to any one of claims 1 to 9 when executed by a processor.

Images