Information recommendation method, device, and computer-readable storage medium

Abstract

This invention discloses an information recommendation method, device, and computer-readable storage medium. The method includes: acquiring customer group screening information corresponding to multimedia information and user communication data of at least one information recipient; determining target information recipients from among the information recipients whose user communication data matches the customer group screening information; determining at least one delivery area based on the user communication location of each target information recipient; determining user group preference type scores for each delivery area based on the customer group screening information and the user communication data of each target information recipient; generating a target user heat map based on each delivery area and the user group preference type scores for each delivery area; and sending the target user heat map to the information recommender. This invention achieves targeted delivery of information through offline channels, improving the effectiveness of information recommendation.

Description

Technical Field

[0001] This invention relates to the field of big data technology, and in particular to an information recommendation method, device, and computer-readable storage medium. Background Technology

[0002] With the continuous progress of the Internet era and the further improvement of network infrastructure, domestic new media has rapidly improved in terms of user numbers, industry scale, and service quality. Multimedia information has permeated everyone's daily life. In the process of building a massive new media platform, a large amount of high-quality content has emerged, and some of this high-quality content has an extremely strong demand for promotion. With all kinds of multimedia information emerging one after another, only by accurately delivering multimedia information to the target audience who truly have the needs can the value of multimedia information be realized.

[0003] In related technologies, the matching of users with multimedia information is usually achieved by passively receiving push requests sent by the user's terminal, or by analyzing the access data based on the user's passive access to obtain recommended videos or information suitable for the user. For example, short video platforms recommend videos that users are interested in in this way. However, this method is too passive and cannot achieve proactive targeting of multimedia information, so it cannot guide users and the recommendation effect of multimedia information is poor. Summary of the Invention

[0004] The main objective of this invention is to provide an information recommendation method, device, and computer-readable storage medium, which aims to screen information recipients through user communication data, proactively obtain information about target information recipients, thereby enabling the targeted delivery of information to offline channels and improving the effectiveness of information recommendation.

[0005] To achieve the above objectives, the present invention provides an information recommendation method, the method comprising the following steps:

[0006] Acquire customer segmentation information corresponding to multimedia information and user communication data of at least one information recipient;

[0007] From among the aforementioned information recipients, determine the target information recipients whose user communication data matches the customer group screening information;

[0008] At least one target area is determined based on the user communication location of each target information recipient, and the user group preference type score of each target information recipient is determined based on the customer group screening information and the user communication data of each target information recipient.

[0009] Based on the target areas and the user group preference type scores of each target area, a target user heat map is generated, and the target user heat map is sent to the information recommender so that the information recommender can make offline information recommendations based on the target user heat map.

[0010] To achieve the above objectives, the present invention also provides an information recommendation device, the device comprising:

[0011] The acquisition module is used to acquire customer group filtering information corresponding to multimedia information and user communication data of at least one information recipient.

[0012] The matching module is used to determine the target information recipients from among the information recipients whose user communication data matches the customer group screening information;

[0013] The determination module is used to determine at least one delivery area based on the user communication location of each target information recipient, and to determine the user group preference type score of each delivery area based on the customer group screening information and the user communication data of each target information recipient.

[0014] The first recommendation module is used to generate a target user heat map based on each of the areas to be targeted and the user group preference type scores of each of the areas to be targeted, and to send the target user heat map to the information recommender so that the information recommender can make offline information recommendations based on the target user heat map.

[0015] To achieve the above objectives, the present invention also provides an information recommendation device, the information recommendation device comprising: a memory, a processor, and an information recommendation program stored in the memory and executable on the processor, wherein the information recommendation program, when executed by the processor, implements the steps of the information recommendation method as described above.

[0016] Furthermore, to achieve the above objectives, the present invention also proposes a computer-readable storage medium storing an information recommendation program, which, when executed by a processor, implements the steps of the information recommendation method described above.

[0017] In this embodiment of the invention, by acquiring customer screening information corresponding to multimedia information and user communication data of at least one information recipient, the information recommender obtains customer screening information for the target audience of the multimedia information and user communication data of the information recipient. The customer screening information reflects the characteristics of the multimedia information and the information recommender's needs and preferences for the target audience of the multimedia information. The user communication data reflects the communication behavior of the information recipient, which can better reflect the information recipient's needs and preferences for the multimedia information. Furthermore, by determining the target information recipient whose user communication data matches the customer screening information from among the information recipients, the matching of the needs and preferences of both the information recommender and the information recipient is achieved. In this way, the selected target information recipient can obtain multimedia information that meets their own needs and preferences. Media information can satisfy the needs and preferences of information recommenders, thus enabling precise matching of multimedia information with target audiences. Furthermore, by determining at least one delivery area based on the communication locations of the target information recipients, and by determining the user group preference type score for each delivery area based on the audience filtering information and the communication data of the target information recipients, the target audience is regionalized, and the user group preference type score for each delivery area is determined. Then, a target user heat map is generated based on each delivery area and its user group preference type score, and this heat map is sent to the information recommender for offline information recommendations, providing a direct display of the target audience for offline information recommendations. In this way, the information recommender only needs to provide multimedia information and audience filtering information to intuitively and conveniently understand the distribution of the target audience for the multimedia information from the target user heat map, thereby enabling proactive targeted delivery. Compared to passively obtaining customer information based on users' spontaneous needs or preferences and their actions, this application proactively guides users before they take any action. This not only provides information recipients with multimedia information that aligns with their needs and preferences, offering convenience, but also improves the recommendation effectiveness of multimedia information, fulfilling the promotional needs of the recommender. It solves the technical problem of poor recommendation effectiveness in multimedia information due to overly passive recommendation methods in related technologies. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating an embodiment of the information recommendation method of the present invention;

[0019] Figure 2 This is a schematic diagram of the information recommendation system in an embodiment of the present invention;

[0020] Figure 3 This is a flowchart illustrating an embodiment of step S20 of the information recommendation method of the present invention;

[0021] Figure 4 This is a schematic diagram of the structure of an embodiment of the information recommendation device of the present invention;

[0022] Figure 5 This is a schematic diagram of the hardware operating environment involved in the embodiments of the present invention.

[0023] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0024] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0025] Reference Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the information recommendation method of the present invention.

[0026] This invention provides an embodiment of an information recommendation method. It should be noted that although the logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order. In this embodiment, the executing entity of the information recommendation method can be an information recommendation device, an information recommendation terminal device, or a server. This embodiment uses an information recommendation device as an example. This information recommendation device can be integrated into devices such as smartphones, personal computers, and servers with data processing capabilities, and is not limited in this embodiment. In this embodiment, for ease of description, the executing entity is omitted. In this embodiment, the information recommendation method includes the following steps:

[0027] Step S10: Obtain customer group screening information corresponding to the multimedia information and user communication data of at least one information recipient;

[0028] In this embodiment, information recommendation involves an information recommender and an information receiver. Information recommendation refers to the process by which the information recommender recommends multimedia information to the information receiver. The multimedia information refers to information disseminated through multimedia formats, such as articles on public accounts, short videos, web page images, offline stalls, physical light boards, and offline performances. Exemplarily, information recommendation can also involve an information distributor. The information distributor can help the information recommender identify the target audience and recommend multimedia information to that audience. Exemplarily, the multimedia information can be an advertisement, the information recommender can be an advertiser, and the information distributor can be a promotional platform. This information recommendation method can be applied to both information recommenders and information distributors.

[0029] The customer screening information refers to information used to screen recipients of multimedia information. It can be determined through statistical analysis of business data or through machine learning models; this embodiment does not impose any limitations on this. The customer screening information may include at least one of the following: information content tags, customer profile tags, customer location information, and customer activity time. The information content tags refer to feature information used to characterize the content of the multimedia information. These can be keywords contained within the multimedia information content, or words representing at least part of the meaning of the multimedia information, such as education, food, maternal and infant products, shopping, sports and fitness, automobiles, beauty, finance, tourism, real estate, entertainment, home decoration, lifestyle, games, horoscopes, wedding dresses, pets, and emotions. The information content tags can be extracted from the multimedia information through methods such as text extraction, speech recognition, and image recognition. For example, the information content tag for a car video's multimedia information could be "car..." The text describes a multimedia message in a WeChat public account's post announcing a live stream. The message includes information on cosmetics and food, and the tags for this multimedia message can include live stream, makeup, and food. The customer profile tags refer to the characteristics of the target customer group for the multimedia message, such as age, residence, and identity. The customer location information refers to the geographical location of the target customer group, which can be the location of communication base stations in that area. The communication industry refers to providers of communication and internet services, such as China Unicom, China Mobile, and China Telecom. The location information provider can determine the customer location using a map or by inputting the communication base station ID (Identity Document). The customer activity time refers to the time it takes for the target customer group to perform a conversion action based on any multimedia message.

[0030] The user communication data refers to user communication data obtained from the telecommunications industry, including at least one of the following: user communication number, internet communication log, user basic attribute data, and user communication location. The user communication number can be the user's registered mobile phone number; the internet communication log can be the mobile internet access log generated by the user's registered mobile phone number; the user communication location can be obtained through the supplier's base station data; the user basic attribute data can be obtained through the telecommunications industry, such as age, gender, place of residence, identity, etc.

[0031] In one feasible implementation, after obtaining the multimedia information to be recommended sent by the information recommender, customer group screening information can be extracted from the multimedia information to be recommended, and / or customer group screening information can be obtained from the information recommender, and some or all users in the communications industry can be identified as information recipients, and user communication data of each of the information recipients can be obtained from the communications industry.

[0032] In one feasible approach, the multimedia information to be recommended can be a set of multimedia information with the same theme but different recommendation formats. For example, for a product, it can be recommended simultaneously through multiple formats such as WeChat official account articles, short videos, SMS messages, and offline light boards. In this case, each type of multimedia information in the multimedia information set can be matched with a target information recipient separately, or a set of multimedia information can be uniformly matched with the same target information recipient.

[0033] Step S20: Determine the target information recipients from among the information recipients whose user communication data matches the customer group screening information;

[0034] In one feasible implementation, the customer group screening information is matched with the user communication data corresponding to each information recipient. The information recipient whose user communication data matches the customer group screening information is determined as the target information recipient. The method of matching the customer group screening information with the user communication data corresponding to each information recipient can be to extract keywords from the user communication data corresponding to each information recipient and perform similarity matching with the customer group screening information, or to convert the user communication data corresponding to each information recipient and the customer group screening information into vectors and then perform matching by inputting a preset matching model. This embodiment does not limit this method.

[0035] For example, customer screening information can be packaged into a customer screening request and sent to the server. The server parses the customer screening request and concatenates it into SQL search conditions. Based on the SQL search conditions, it searches in the database that stores user communication data and retrieves users who meet the SQL search conditions, storing them in the target information recipient table.

[0036] Optionally, the step of determining the target information recipient from among the information recipients that matches the user communication data with the customer group screening information includes:

[0037] Step A10: Determine the initial information recipients from among the information recipients whose user communication data matches the customer group screening information;

[0038] In this embodiment, when the information recommender determines that multimedia information will be recommended through offline channels, information recipients who prefer to scan QR codes for offline information are more likely to convert to the multimedia information.

[0039] In one feasible implementation, the customer group screening information is matched with the user communication data corresponding to each information recipient, and the information recipient whose user communication data matches the customer group screening information is determined as the initial information recipient.

[0040] Step A20: Check whether each of the initially selected information recipients has a historical scanning record;

[0041] Step A30: Identify the initial information recipients with historical scanning records as the target information recipients.

[0042] In one feasible implementation, the scanning information of each information recipient can be recorded and stored as historical scanning records after each acquisition of user communication data from the information recipient. After determining the target information recipient, it can be first queried whether each of the initially selected information recipients has historical scanning records, and the initially selected information recipients with historical scanning records can be determined as the target information recipients.

[0043] Optionally, before the step of determining at least one delivery area based on the user communication location of each of the target information recipients, the method further includes:

[0044] The QR code location information in the historical scanning records corresponding to each of the target information recipients is determined as the user communication location corresponding to each of the target information recipients.

[0045] In one feasible implementation, the historical scanning records may include QR code location information. Thus, after identifying the target information recipient with historical scanning records, the QR code location information can be extracted from the historical scanning records of each target information recipient. This allows the determination of the QR code location information corresponding to each target information recipient. The QR code location information corresponding to each target information recipient is the location where each target information recipient was when scanning the QR code. Therefore, the QR code location information corresponding to each target information recipient can be determined as the user communication location corresponding to each target information recipient.

[0046] Optionally, after the step of determining the target information recipient from each of the information recipients that matches the user communication data with the customer group screening information, the method further includes:

[0047] Information recommendations are made to the target information recipients.

[0048] In one feasible implementation, after determining the target information recipients, information can be recommended to each of the target information recipients. The information recommendation methods include sending text messages, pushing articles from public accounts, and pushing application messages.

[0049] For example, for online customers, after identifying the target recipients of the information, lead generation text messages can be sent at regular intervals and in fixed quantities.

[0050] For example, for offline customer groups, potential target information recipients can be identified from each of the information recipients whose user behavior preferences match the information content tags and whose user basic attribute data matches the customer profile tags. Then, the user communication location of each potential target information recipient can be continuously or intermittently obtained. If the user communication location matches the customer group location information, the potential target information recipient whose user communication location matches the customer group location information is identified as the target information recipient. Then, a referral SMS can be proactively pushed to the target information recipient. This can achieve timely and accurate push of referral SMS to the target customer group when the target customer group arrives at the coverage area of ​​the set customer group location information.

[0051] For example, during the information recommendation process, the information recipients' data can be continuously accumulated to provide historical data references for subsequent information recommendations. Information recommendation evaluations can also be conducted on information recipients, and must-send lists, exclusion lists, blacklists, whitelists, etc. can be established, with different information recommendation rules formulated for information recipients on each list.

[0052] Optionally, the step of determining the target information recipient from among the information recipients that matches the user communication data with the customer group screening information includes:

[0053] Step B10: Determine the initial information recipients from among the information recipients whose user communication data matches the customer group screening information;

[0054] Step B20: Obtain the current delivery frequency of each of the initially selected information recipients;

[0055] Step B30: Determine the target information receiver whose current delivery frequency meets the preset frequency condition from among the initially selected information receivers.

[0056] In this embodiment, for information recommendations made through online channels, a low frequency of recommendations may fail to attract the attention of the information recipient, resulting in poor recommendation effectiveness. However, a high frequency may easily cause resentment among the information recipient. Therefore, frequency conditions can be set to further filter the target information recipients. The frequency conditions can be set according to actual circumstances, and this embodiment does not impose any restrictions on them.

[0057] In one feasible implementation, the customer group screening information is matched with the user communication data corresponding to each information recipient. Information recipients whose user communication data matches the customer group screening information are identified as initial information recipients. After identifying the initial information recipients, the current delivery frequency of each initial information recipient can be obtained. The current delivery frequency of each initial information recipient is compared with a preset frequency condition. Initial information recipients that meet the preset frequency condition are identified as target information recipients. For example, the preset frequency condition can be once every n days. If the time x is determined to be the latest time x when the initial information recipient recommended the multimedia information, then the current delivery frequency of the initial information recipient is determined to be once every x days. If x is less than n, then the current delivery frequency of the initial information recipient does not meet the preset frequency condition. If x is greater than or equal to n, then the initial information recipient is identified as a target information recipient.

[0058] Step S30: Determine at least one target area based on the user communication location of each target information recipient, and determine the user group preference type score of each target information recipient based on the customer group screening information and the user communication data of each target information recipient.

[0059] In this embodiment, the influence of information recommendations through offline channels has certain regional limitations. For example, the probability of information recipients who do not take public transportation seeing multimedia information displayed on signs at bus stops is relatively low. Therefore, by sending the target information recipients' communication locations to the information recommender, the recommender can select areas with a larger target customer base for offline information recommendations. The user communication location can be the location of a communication base station, which can be determined through communication data generated by the user's terminal; or it can be the location where the user is when scanning a QR code, which can be determined by detecting the location information contained in the scanned QR code. However, selecting locations based on the scattered or listed communication locations of all target information recipients requires the information recommender to manually compare and select all the communication locations of all target information recipients, which is time-consuming and labor-intensive, and can only provide a rough estimate with low accuracy.

[0060] The user group preference type score refers to the overall score of a user behavior preference among information recipients within a certain area. For example, each user behavior preference can be scored in advance for some or all information recipients in each area to be targeted, resulting in a user behavior preference score for each user behavior preference of each information recipient in each area to be targeted. Then, the user group preference type score for each user behavior preference in each area to be targeted can be calculated by averaging, summing, weighted averaging, weighted summing, etc.

[0061] In one feasible implementation, one or more target information receiving areas containing at least one target information receiving area are determined based on the user communication location of each target information receiving area. Based on the customer group screening information, the user communication data of each target information receiving area, and preset scoring conditions, each target information receiving area is scored to determine the user behavior preference score of each target information receiving area. Then, the user behavior preference scores of the target information receiving areas corresponding to each target information receiving area are statistically analyzed to determine the user group preference type score of each target information receiving area.

[0062] In a specific implementation, the area to be delivered can be divided based on the user communication location of each target information recipient, combined with the estimated impact range of the multimedia information, the distribution of buildings or facilities, and other information. Alternatively, at least one delivery area can be pre-defined, and the delivery area containing the user communication location of at least one target information recipient can be determined as the delivery area. The specific determination can be made according to the actual situation, and this embodiment does not impose any restrictions on this.

[0063] Optionally, the user communication data includes internet communication logs, and the step of determining the user group preference type score for each of the target information recipients based on the customer group screening information and the user communication data of each of the target information recipients includes:

[0064] Step S31: Determine the user behavior preferences and information conversion channel scores of each target information recipient based on the Internet communication logs;

[0065] In this embodiment, after determining the target information recipients, the user group preference type score corresponding to the information content tag of each target information recipient can be determined based on the user communication data of each target information recipient. The user group preference type score determined in this way is more consistent with the current multimedia information, which can further filter the target information recipients and improve the accuracy of site selection.

[0066] The information conversion channels include QR code scanning, SMS, phone calls, official accounts, and application messages. The information conversion channel score refers to the score reflecting the user behavior preferences of each target information recipient across various information conversion channels. This score can be determined by analyzing user actions such as clicks, redirects, browsing, and purchases generated through these channels. For example, SMS operation data such as click counts, browsing time, and purchase amount can be statistically analyzed. This data can then be normalized, weighted averaged, and / or weighted summed to obtain the SMS operation score. Specific statistical analysis methods for the information conversion channel score can refer to existing technologies or be determined based on actual circumstances; this embodiment does not impose any limitations on this. The user operation score reflects the operational preferences of the information recipient. For example, some information recipients are more receptive to recommendations from official accounts, while others are more receptive to recommendations from physical advertisements.

[0067] In one feasible implementation, after obtaining user communication data from each of the information recipients from the communications industry, internet communication logs can be obtained from the user communication data. Furthermore, the internet communication logs can be analyzed. User behavior preferences matching the internet communication logs can be determined from preset preference types using machine learning models, keyword extraction, similarity matching, and other methods. Conversion data for various information conversion channels can also be queried from the internet communication logs, and statistical analysis can be performed on the conversion data to determine the information conversion channel score.

[0068] Step S32: Determine the user behavior preference score corresponding to each of the user behavior preferences of each of the target information recipients based on the user behavior preferences and information conversion channel scores of each of the target information recipients.

[0069] In one feasible implementation, the user behavior preferences and information conversion channel scores of each target information recipient are analyzed through algorithms, models, statistical analysis, etc., to determine the user behavior preference scores corresponding to each user behavior preference of each target information recipient.

[0070] Optionally, the step of determining the user behavior preference score corresponding to each of the user behavior preferences of each of the target information recipients based on the user behavior preferences of each of the target information recipients and the information conversion channel score includes:

[0071] The user behavior preferences and information conversion channel scores of each target information recipient are input into a preset user behavior scenario preference algorithm to obtain the user behavior preference scores corresponding to each user behavior preference of each target information recipient. The user behavior scenario preference algorithm is as follows:

[0072]

[0073] Where, p i,j Let x be the user behavior preference score of user j for information receiver i. i,k For information recipient i, the information conversion channel score for information conversion channel k is w. i,j The weights of user behavior preferences j for the pre-defined information receiver i.

[0074] Step S33: Determine the user group preference type score for each of the regions to be targeted based on the user behavior preference score corresponding to each region.

[0075] In one feasible implementation, statistical analyses such as averaging, summing, weighted averaging, and weighted summing are performed on the user behavior preference scores corresponding to each of the said target areas to determine the user group preference type scores of each of the said target areas.

[0076] Step S40: Generate a target user heat map based on each of the areas to be targeted and the user group preference type scores of each of the areas to be targeted, and send the target user heat map to the information recommender so that the information recommender can make offline information recommendations based on the target user heat map.

[0077] In this embodiment, the information recommendation channels include online and offline channels. Offline channels are more intuitive and have a longer recommendation duration, giving them certain advantages. However, the choice of location for offline channels has a significant impact on the effectiveness of information recommendation. For example, for multimedia information related to home renovation, offline channels are more effective in areas such as newly opened residential communities and newly delivered residential communities; for multimedia information related to tourism, offline channels are more effective in areas such as tourist attractions, train stations, and airports.

[0078] When there are multiple regions to be targeted, the difference in the number and / or quality of the target customer group in each region may be small. Information recommenders can combine other factors such as the number of media available for information delivery, delivery format, and specific site conditions when selecting a location. However, because the difference in the number and / or quality of the target customer group is small, directly sending messages or displaying the user's communication location on a map is not intuitive enough. By generating a target user heat map, the number and quality of the target customer group in each region to be targeted can be displayed more intuitively.

[0079] The target user heat map refers to a map that can display the user group preference type scores of each area in the map, so that the information recommender can intuitively see the user group preference type scores of each area to be advertised. For example, the locations and / or contact information of elevator advertising space partners, bus stop electronic advertising space partners, vacant shops, and community property management companies within a certain distance of the communication location of each target information recipient can also be output and displayed on the target user heat map.

[0080] In one feasible implementation, a target user heat map is generated based on the user group preference type scores of each of the areas to be targeted. This target user heat map is then sent to the information recommender, who can select suitable information recommendation locations for offline information recommendations based on the heat map. This saves the information recommender time and effort in site selection, and the quantified scores and clear regional divisions effectively improve the accuracy of site selection for the areas to be targeted.

[0081] For example, after the step of sending the target user heat map to the information recommender, feedback information from each of the target information recipients can be obtained, and information recommendation scores can be calculated based on the feedback information to estimate and predict the marketing effect of the information recommender's information recommendation. This can also provide a reference for adjusting various preset parameters in the information recommendation method.

[0082] Optionally, after the step of sending the target user heat map to the information recommender, the method further includes:

[0083] Step S50: Obtain feedback information from each of the target information recipients;

[0084] In this embodiment, the feedback information refers to information transformation-related information generated by each of the target information recipients after receiving multimedia information, such as click operations, browsing operations, purchase operations, and other data generated by each of the target information recipients based on the multimedia information. For example, user number fronthaul technology can be applied (when a mobile user accesses mobile internet services via WAP (Wireless Application Protocol) and APN (Access Point Name), the WAP gateway obtains the user's MSISDN (Mobile Subscriber International ISDN / PSTN number) and private network IP address through RADIUS (Remote Application Dialing User Service) messages, and transmits the user number to the page for user identification by inserting the user number into the HTTP (Hypertext Transfer Protocol) header). This records user identity and content dwell data (access time, accessed content, whether a button was clicked, button location, dwell time, etc.). By accessing content dwell data and the content itself (video duration, subtitle location, video / image tone, video / image resolution, video / image contrast, video / image brightness, QR code location, button location), the system can record user identity and content dwell data (access time, accessed content, whether a button was clicked, button location, dwell time, etc.).

[0085] In one feasible implementation, after sending the user communication location of each of the target information recipients to the information recommender, feedback information from each of the target information recipients can also be obtained.

[0086] Step S60: Iteratively update the weights of each user behavior preference of each target information recipient based on the feedback information.

[0087] In one feasible implementation, the weights of the user behavior preferences of each target information recipient are iteratively updated based on the feedback information to provide more data for subsequent multimedia information recommendations. For example, if a user whose car behavior preference weight is 50% clicks on the push details of the "car" content in the multimedia information, then the weight of that user's car behavior preference can be increased. The specific increase can be determined based on feedback information such as clicks and browsing time. As another example, if a user whose gasoline car behavior preference weight is 90% clicks on the push details of the "new energy" content in the multimedia information, then the weight of that user's gasoline car behavior preference can be decreased. The specific decrease can be determined based on feedback information such as clicks and browsing time. In this way, the delivery of subsequent multimedia information related to "gasoline cars" can be reduced, while the delivery of multimedia information related to "new energy" can be increased.

[0088] Optionally, the feedback information includes feedback behavior preferences, and the step of iteratively updating the weights of each user behavior preference of each target information recipient based on the feedback information includes:

[0089] Step S61: Based on the correlation between preset behavioral preferences, determine the correlation between the feedback behavioral preferences of each target information receiver and their corresponding user behavioral preferences.

[0090] Step S62: Decrease the weight of user behavior preferences corresponding to negative correlation, and increase the weight of user behavior preferences corresponding to positive correlation.

[0091] In this embodiment, there may be certain correlations between various behavioral preferences. These correlations include positive and negative correlations. For example, banks and finance, cosmetics and clothing usually have a certain positive correlation, while food and weight loss, new energy and coal usually have a certain negative correlation. Therefore, when adjusting the weights, the correlations between various behavioral preferences can be considered to improve the effectiveness of the information.

[0092] In one feasible implementation, the correlation between various behavioral preferences can be preset, and then, based on the preset correlation between behavioral preferences, it can be determined whether each user behavioral preference of each target information recipient is related to its feedback behavioral preference. Then, the weight of user behavioral preferences with negative correlation is reduced, and the weight of user behavioral preferences with positive correlation is increased.

[0093] In one feasible approach, the information recommendation method can be applied to an information recommendation system, which refers to... Figure 2The system includes at least a multimedia acquisition system, a communication data acquisition system, a customer group screening condition setting system, a customer group matching system, a screening system, a delivery system, and a feedback system. The multimedia acquisition system includes a multimedia information acquisition module and an information content tag setting module. The multimedia information acquisition module is used to acquire multimedia information through multiple channels; the information content tag setting module is used to set information content tags for the multimedia information. The communication data acquisition system includes a communication data acquisition module and a data classification module. The communication data acquisition module is used to acquire user communication data from communication agent service providers; the data classification module is used to classify the user communication data. The customer group screening condition setting system includes a screening condition setting tool, which allows users to set various screening conditions. The customer group matching system includes an online matching module and an offline matching module. The online matching module is used to match online users and multimedia information; the offline matching module is used to match offline users and multimedia information. The screening system includes a screening tool, which is used to filter target customers from the matched potential target customer groups. The delivery system includes an online delivery module and an offline delivery module. The online delivery module delivers multimedia information to the matched target audience online. The offline delivery module delivers multimedia information to the matched target audience offline using either a first mode or a second mode. The feedback system includes a feedback scoring module and an attractiveness scoring module. The feedback scoring module tracks user feedback and adjusts the corresponding user's preference weight based on the feedback information, where the preference weight affects the user's probability of successful matching. The attractiveness scoring module is used to accumulate data on the effectiveness of the content activities delivered by the system. The system assigns an attractiveness score and estimates the delivery effect for newly uploaded content.

[0094] In this embodiment, by acquiring customer screening information corresponding to multimedia information and user communication data of at least one information recipient, the information recommender obtains customer screening information for the multimedia information delivery target audience and user communication data of the information recipient. The customer screening information reflects the characteristics of the multimedia information and the information recommender's needs and preferences for the multimedia information delivery target audience. The user communication data reflects the communication behavior of the information recipient, which can better reflect the information recipient's needs and preferences for multimedia information. Furthermore, by determining the target information recipient whose user communication data matches the customer screening information from among the information recipients, the matching of needs and preferences between the information recommender and the information recipient is achieved. Thus, the selected target information recipient can obtain multimedia information that meets their own needs and preferences. This approach combines multimedia information with the needs and preferences of information recommenders, enabling precise matching of multimedia information to target audiences. Furthermore, by determining at least one delivery area based on the communication locations of the target information recipients, and by determining user group preference type scores for each delivery area based on the customer group filtering information and the communication data of the target information recipients, the target audience is regionalized, and a user group preference type score is determined for each delivery area. Then, a target user heat map is generated based on each delivery area and its user group preference type score, and this heat map is sent to the information recommender for offline information recommendations. This provides a direct visualization of the target audience for offline information recommendations. In this way, information recommenders only need to provide multimedia information and customer group filtering information to intuitively and conveniently understand the distribution of the target audience for the multimedia information from the target user heat map, thus enabling proactive targeted delivery. Compared to passively obtaining customer information based on users' spontaneous needs or preferences and their actions, this application proactively guides users before they take any action. This not only provides information recipients with multimedia information that aligns with their needs and preferences, offering convenience, but also improves the recommendation effectiveness of multimedia information, fulfilling the promotional needs of the recommender. It solves the technical problem of poor recommendation effectiveness in multimedia information due to overly passive recommendation methods in related technologies.

[0095] Furthermore, based on the above embodiments, another embodiment of the information recommendation method of the present invention is proposed, with reference to... Figure 3 , Figure 3 This is a flowchart illustrating an embodiment of step S20 of the information recommendation method of the present invention. In this embodiment, the customer group filtering information includes at least one of information content tags, customer group profile tags, and customer group location information.

[0096] The user communication data includes at least one of internet communication logs, user basic attribute data, and user communication location.

[0097] Step S20 includes:

[0098] Step C10: Determine the user behavior preferences of each information recipient based on the Internet communication logs;

[0099] In this embodiment, the target customer group for information recommendation includes online and offline customers. The online customer group refers to potential customers recommended through online channels, such as app messages, WeChat official account articles, SMS, and telephone calls. The offline customer group refers to potential customers recommended through offline channels, such as physical stalls, signs in elevators / hallways / subways / bus stops, and offline performances. Different user communication data can be used to specifically filter different target customer groups, improving the accuracy of target customer group selection.

[0100] In one feasible implementation, after obtaining user communication data from each of the information recipients from the telecommunications industry, internet communication logs can be obtained from the user communication data. Furthermore, the internet communication logs can be analyzed. For example, user behavior preferences matching the internet communication logs can be determined from preset preference types using machine learning models, keyword extraction, similarity matching, etc. The user behavior preferences in this solution include one or more of the following: education, food, maternal and infant care, shopping, sports and fitness, automobiles, beauty, finance, travel, real estate, entertainment, home decoration, lifestyle, games, horoscopes, wedding dresses, pets, and emotions. For example, if a user's internet communication logs show that the user purchases pet supplies or browses pet websites more frequently than a preset frequency threshold, then the user's internet behavior preference can be determined to be "pets."

[0101] Step C20: Determine at least one target information receiver from among the information receivers that meets the customer group matching conditions, wherein the customer group matching conditions include at least one of the following: the user's behavioral preferences match the information content tags, the user's basic attribute data matches the customer group profile tags, and the user's communication location matches the customer group location information.

[0102] In one feasible implementation, it is determined sequentially whether the user communication data corresponding to each information recipient meets the customer group matching conditions, and at least one information recipient that meets the customer group matching conditions is determined as the target information recipient. The customer group matching conditions include at least one of the following: the user behavior preferences match the information content tags, the user basic attribute data matches the customer profile tags, and the user communication location matches the customer location information. In this context, matching user behavior preferences with information content tags can mean that the user behavior preferences are the same as the information content tags, that the user behavior preferences include the information content tags, or that a portion of the user behavior preferences has a similarity higher than a preset similarity threshold. Matching user basic attribute data with customer profile tags can mean that the user basic attribute data is the same as the customer profile tags, that the user basic attribute data includes the customer profile tags, or that a portion of the user basic attribute data has a similarity higher than a preset similarity threshold. Matching user communication location with customer location information can mean that the user communication location is within the coverage area of ​​the customer location information, that the user communication location enters the coverage area of ​​the customer location information, or that the user communication location stays within the coverage area of ​​the customer location information for a time exceeding a preset time threshold. These specific settings can be implemented according to actual circumstances, and this embodiment does not impose any limitations on them.

[0103] For example, for online customer groups, the customer group matching criteria can be that the user's behavioral preferences match the information content tags and the user's basic attribute data matches the customer group profile tags. For instance, a first user can be determined by matching the user's behavioral preferences with the information content tags, and a second user can be determined by matching the user's basic attribute data with the customer group profile tags. Users who are the same as the first user and the second user can be identified as target information recipients.

[0104] For example, for offline customer groups, the customer group matching conditions can be that the user's behavioral preferences match the information content tags, the user's basic attribute data matches the customer group profile tags, and the user's communication location matches the customer group location information. For instance, a first user can be determined by matching the user's behavioral preferences with the information content tags; a second user can be determined by matching the user's basic attribute data with the customer group profile tags; and a third user can be determined by matching the user's communication location with the customer group location information. Users who are the same among the third user, the second user, and the first user are then identified as target information recipients. Since user communication locations may change in real time, potential target information recipients whose user behavioral preferences match the information content tags and whose user basic attribute data matches the customer group profile tags can be identified from among the information recipients. Then, the communication locations of each potential target information recipient can be continuously or intermittently acquired. If a user's communication location matches the customer group location information, the potential target information recipient whose communication location matches the customer group location information is identified as the target information recipient.

[0105] Optionally, the steps for obtaining information content tags include:

[0106] Step D10: Obtain multimedia information by crawling the associated information corresponding to the multimedia information.

[0107] In this embodiment, the multimedia information is usually concise and carries limited information. By using web crawlers, more related information can be obtained from the Internet without the information recommender having to upload each piece of information. This not only provides convenience for the information recommender but also improves the comprehensiveness of the information content tags.

[0108] In one feasible implementation, the multimedia information to be recommended can be obtained directly from the message recommender or from the URL sent by the message recommender. The associated information corresponding to the multimedia information can then be obtained through web crawling. The associated information includes author, title, text, cover image, illustrations, tags, etc.

[0109] For example, when the multimedia information is video information, the video author, video title, video text, and video file can be obtained via web crawler as associated information corresponding to the multimedia information. In one feasible approach, to facilitate the information recommender in viewing the general content of the video on the page after uploading the multimedia information without placing too much pressure on the website, partial frames of the video file can be retrieved to create a GIF file, which can then be displayed for the information recommender to select.

[0110] For example, when the multimedia information is text information or image information, a web crawler can obtain the author, title, text, and text or image file of the text or image as associated information corresponding to the multimedia information.

[0111] For example, the information recommender can input the multimedia URL link corresponding to the multimedia information. For instance, if the multimedia URL link is the URL link of a public account article, the public account account (author), public account article title, text, video file (MP4 format), cover image file, and image files within the article can be crawled from the multimedia URL link.

[0112] Step D20: Determine the information content tags of the multimedia information based on the associated information.

[0113] In one feasible implementation, the information content tags of the multimedia information can be determined by inputting the associated information into a trained machine learning classification model. The machine learning classification model is similar to existing technologies and will not be described in detail here. Alternatively, information content tags that match the associated information can be determined from preset classification tags by keyword matching.

[0114] Optionally, the step of determining the information content tag of the multimedia information based on the association information includes:

[0115] Step D21: Match the associated information with a preset keyword database to determine at least one target keyword corresponding to the multimedia information;

[0116] Step D22: Determine the target category label corresponding to each target keyword based on the preset mapping relationship between keywords and category labels;

[0117] Step D23: Count the number of times the target category label is determined, and determine the target category label whose determination frequency meets the preset determination frequency condition as the information content label of the multimedia information.

[0118] In one feasible implementation, a keyword lexicon can be pre-established, and a mapping relationship between each keyword in the keyword lexicon and a category tag can be established. Then, the associated information can be matched with the pre-established keyword lexicon, and one or more target keywords that have appeared in the associated information or whose similarity to some sentences in the associated information is higher than a pre-established similarity threshold can be selected from the pre-established keyword lexicon. Then, based on each target keyword, the mapping relationship between the pre-established keywords and category tags can be queried to determine the target category tag corresponding to each target keyword. The number of times the target category tag is determined is counted. After the target category tags corresponding to all target keywords are determined, the target category tag whose number of determinations meets the pre-established determination number condition is determined as the information content tag of the multimedia information. The pre-established determination number condition can be a target category tag that exceeds the pre-established determination number threshold, the target category tag with the most determinations, the target category tag that exceeds the pre-established determination number percentage threshold, or the target category tag with the highest determination number percentage, etc.

[0119] In this embodiment, internet communication logs can characterize user online behavior, reflecting user needs and preferences. Target information recipients selected based on internet communication logs have needs and preferences that better match the content of multimedia information. In other words, target information recipients selected based on internet communication logs are more likely to be interested in the multimedia information content, resulting in higher acceptance and better recommendation effectiveness. Target customer groups with needs or preferences for multimedia information often share certain commonalities. For example, women have a greater need and preference for cosmetics, and teachers have a greater need and preference for throat lozenges. Therefore, filtering target customer groups using basic user attribute data can improve the accuracy of target customer group selection. Some multimedia information has obvious regional commonalities. For example, tourism-related multimedia information can be sent to recipients at train stations, airports, and tourist attractions in tourist cities. Filtering based on user communication locations can improve the accuracy of target customer group selection and more accurately determine the location of offline customers, thus helping information recommenders to more accurately determine the offline recommendation location for multimedia information.

[0120] In addition, refer to Figure 4 This invention also proposes an information recommendation device, which includes:

[0121] The acquisition module 10 is used to acquire customer group screening information corresponding to multimedia information and user communication data of at least one information recipient;

[0122] Matching module 20 is used to determine, from each of the information recipients, the target information recipients whose user communication data matches the customer group screening information;

[0123] The determination module 30 is used to determine at least one delivery area based on the user communication location of each target information recipient, and to determine the user group preference type score of each delivery area based on the customer group screening information and the user communication data of each target information recipient.

[0124] The first recommendation module 40 is used to generate a target user heat map based on each of the areas to be targeted and the user group preference type scores of each of the areas to be targeted, and send the target user heat map to the information recommender so that the information recommender can make offline information recommendations based on the target user heat map.

[0125] Optionally, the matching module 20 is further configured to:

[0126] From the various information recipients, a preliminary selection of information recipients is determined to match the user communication data with the customer group screening information;

[0127] Check whether each of the aforementioned initial selection information recipients has a historical scanning record;

[0128] The initial information recipients with historical scanning records are identified as the target information recipients.

[0129] Optionally, the determining module 30 is further configured to:

[0130] Based on the internet communication logs, determine the user behavior preferences and information conversion channel scores of each of the target information recipients;

[0131] The user behavior preference score for each of the target information recipients is determined based on the user behavior preferences of each target information recipient and the information conversion channel score.

[0132] The user group preference type score for each of the regions to be targeted is determined based on the user behavior preference score corresponding to each region.

[0133] Optionally, the determining module 30 is further configured to:

[0134] The user behavior preferences and information conversion channel scores of each target information recipient are input into a preset user behavior scenario preference algorithm to obtain the user behavior preference scores corresponding to each user behavior preference of each target information recipient. The user behavior scenario preference algorithm is as follows:

[0135]

[0136] Where, p i,j Let x be the user behavior preference score of user j for information receiver i. i,k For information recipient i, the information conversion channel score for information conversion channel k is w. i,j The weights of user behavior preferences j for the pre-defined information receiver i.

[0137] Optionally, after the operation of sending the target user heat map to the information recommender, the information recommending device further includes a feedback module, which is used to:

[0138] Obtain feedback information from each of the target information recipients;

[0139] The weights of each user behavior preference of each target information recipient are iteratively updated based on the feedback information.

[0140] Optionally, the feedback module is further configured to:

[0141] Based on the correlation between preset behavioral preferences, determine the correlation between the feedback behavioral preferences of each target information receiver and their corresponding user behavioral preferences;

[0142] The weights of user behavior preferences corresponding to negative correlations are adjusted to be reduced, while the weights of user behavior preferences corresponding to positive correlations are adjusted to be increased.

[0143] Optionally, the customer screening information includes at least one of information content tags, customer profile tags, and customer location information;

[0144] The user communication data includes at least one of internet communication logs, user basic attribute data, and user communication location.

[0145] The matching module 20 is further configured to:

[0146] Determine the user behavior preferences of each information recipient based on the internet communication logs;

[0147] At least one target information recipient that meets the customer group matching conditions is determined from the information recipients, wherein the customer group matching conditions include at least one of the following: the user behavior preferences match the information content tags, the user basic attribute data matches the customer group profile tags, and the user communication location matches the customer group location information.

[0148] Optionally, the acquisition module 10 is further configured to:

[0149] Obtain multimedia information, and retrieve the associated information corresponding to the multimedia information through web crawling;

[0150] The information content tags of the multimedia information are determined based on the associated information.

[0151] Optionally, the acquisition module 10 is further configured to:

[0152] The associated information is matched with a preset keyword database to determine at least one target keyword corresponding to the multimedia information;

[0153] Based on the preset mapping relationship between keywords and category tags, determine the target category tag corresponding to each of the target keywords;

[0154] The number of times the target category label is determined is counted, and the target category label whose number of determinations meets the preset determination number condition is determined as the information content label of the multimedia information.

[0155] Optionally, after determining the target information recipients from among the information recipients whose user communication data matches the customer group screening information, the information recommendation device further includes a second recommendation module, the second recommendation module being used to:

[0156] Information recommendations are made to the target information recipients.

[0157] Optionally, the matching module is further configured to:

[0158] From the various information recipients, a preliminary selection of information recipients is determined to match the user communication data with the customer group screening information;

[0159] Obtain the current delivery frequency of each of the initially selected information recipients;

[0160] From the selected information recipients, determine the target information recipients whose current delivery frequency meets the preset frequency conditions.

[0161] The information recommendation device provided by this invention, employing the information recommendation method described in the above embodiments, solves the technical problem that the passive nature of multimedia information recommendation in related technologies leads to poor recommendation results. Compared with the prior art, the beneficial effects of the information recommendation device provided in this invention are the same as those of the information recommendation method described in the above embodiments, and other technical features in this information recommendation device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0162] Furthermore, embodiments of the present invention also propose an information recommendation device, such as... Figure 5 As shown, Figure 5This is a schematic diagram of the device structure of the hardware operating environment involved in the embodiments of the present invention. It should be noted that the recommended device in the embodiments of the present invention can be a smartphone, a personal computer, a server, etc., and no specific limitation is made here.

[0163] like Figure 5 As shown, the information recommendation device may include: a processor 1001, such as a CPU; a network interface 1004; a user interface 1003; a memory 1005; and a communication bus 1002. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be high-speed RAM or non-volatile memory, such as a disk drive. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0164] Those skilled in the art will understand that Figure 5 The device structure shown does not constitute a limitation on the information recommendation device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0165] like Figure 5 As shown, the memory 1005, as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an information recommendation program. The operating system is a program that manages and controls the device's hardware and software resources, supporting the operation of the information recommendation program and other software or programs. Figure 5 In the device shown, the user interface 1003 is mainly used for data communication with the client; the network interface 1004 is mainly used for establishing a communication connection with the server; and the processor 1001 can be used to call the information recommendation program stored in the memory 1005 and perform the following operations:

[0166] Acquire customer segmentation information corresponding to multimedia information and user communication data of at least one information recipient;

[0167] From among the aforementioned information recipients, determine the target information recipients whose user communication data matches the customer group screening information;

[0168] At least one target area is determined based on the user communication location of each target information recipient, and the user group preference type score of each target information recipient is determined based on the customer group screening information and the user communication data of each target information recipient.

[0169] Based on the target areas and the user group preference type scores of each target area, a target user heat map is generated, and the target user heat map is sent to the information recommender so that the information recommender can make offline information recommendations based on the target user heat map.

[0170] Furthermore, the processor 1001 can also be used to call the information recommendation program stored in the memory 1005 and perform the following operations:

[0171] From the various information recipients, a preliminary selection of information recipients is determined to match the user communication data with the customer group screening information;

[0172] Check whether each of the aforementioned initial selection information recipients has a historical scanning record;

[0173] The initial information recipients with historical scanning records are identified as the target information recipients.

[0174] Furthermore, the processor 1001 can also be used to call the information recommendation program stored in the memory 1005 and perform the following operations:

[0175] Based on the internet communication logs, determine the user behavior preferences and information conversion channel scores of each of the target information recipients;

[0176] The user behavior preference score for each of the target information recipients is determined based on the user behavior preferences of each target information recipient and the information conversion channel score.

[0177] The user group preference type score for each of the regions to be targeted is determined based on the user behavior preference score corresponding to each region.

[0178] Furthermore, the processor 1001 can also be used to call the information recommendation program stored in the memory 1005 and perform the following operations:

[0179] The user behavior preferences and information conversion channel scores of each target information recipient are input into a preset user behavior scenario preference algorithm to obtain the user behavior preference scores corresponding to each user behavior preference of each target information recipient. The user behavior scenario preference algorithm is as follows:

[0180]

[0181] Where, p i,j Let x be the user behavior preference score of user j for information receiver i. i,k For information recipient i, the information conversion channel score for information conversion channel k is w.i,j The weights of user behavior preferences j for the pre-defined information receiver i.

[0182] Furthermore, after sending the target user heat map to the information recommender, the processor 1001 can also call the information recommending program stored in the memory 1005 to perform the following operations:

[0183] Obtain feedback information from each of the target information recipients;

[0184] The weights of each user behavior preference of each target information recipient are iteratively updated based on the feedback information.

[0185] Furthermore, the processor 1001 can also be used to call the information recommendation program stored in the memory 1005 and perform the following operations:

[0186] Based on the correlation between preset behavioral preferences, determine the correlation between the feedback behavioral preferences of each target information receiver and their corresponding user behavioral preferences;

[0187] The weights of user behavior preferences corresponding to negative correlations are adjusted to be reduced, while the weights of user behavior preferences corresponding to positive correlations are adjusted to be increased.

[0188] Furthermore, the customer screening information includes at least one of information content tags, customer profile tags, and customer location information;

[0189] The user communication data includes at least one of internet communication logs, user basic attribute data, and user communication location.

[0190] Processor 1001 can also be used to call the information recommendation program stored in memory 1005 and perform the following operations:

[0191] Determine the user behavior preferences of each information recipient based on the internet communication logs;

[0192] At least one target information recipient that meets the customer group matching conditions is determined from the information recipients, wherein the customer group matching conditions include at least one of the following: the user behavior preferences match the information content tags, the user basic attribute data matches the customer group profile tags, and the user communication location matches the customer group location information.

[0193] Furthermore, the processor 1001 can also be used to call the information recommendation program stored in the memory 1005 and perform the following operations:

[0194] Obtain multimedia information, and retrieve the associated information corresponding to the multimedia information through web crawling;

[0195] The information content tags of the multimedia information are determined based on the associated information.

[0196] Furthermore, the processor 1001 can also be used to call the information recommendation program stored in the memory 1005 and perform the following operations:

[0197] The associated information is matched with a preset keyword database to determine at least one target keyword corresponding to the multimedia information;

[0198] Based on the preset mapping relationship between keywords and category tags, determine the target category tag corresponding to each of the target keywords;

[0199] The number of times the target category label is determined is counted, and the target category label whose number of determinations meets the preset determination number condition is determined as the information content label of the multimedia information.

[0200] Furthermore, after determining the target information recipients from among the information recipients whose user communication data matches the customer group screening information, the processor 1001 can also invoke the information recommendation program stored in the memory 1005 to perform the following operations:

[0201] Information recommendations are made to the target information recipients.

[0202] Furthermore, the processor 1001 can also be used to call the information recommendation program stored in the memory 1005 and perform the following operations:

[0203] From the various information recipients, a preliminary selection of information recipients is determined to match the user communication data with the customer group screening information;

[0204] Obtain the current delivery frequency of each of the initially selected information recipients;

[0205] From the selected information recipients, determine the target information recipients whose current delivery frequency meets the preset frequency conditions.

[0206] The electronic device provided by this invention employs the information recommendation method described in the above embodiments, solving the technical problem that the passive nature of multimedia information recommendation in related technologies leads to poor recommendation performance. Compared with the prior art, the beneficial effects of the electronic device provided by this invention are the same as those of the information recommendation method described in the above embodiments, and other technical features of this electronic device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0207] Furthermore, this invention also proposes a computer-readable storage medium storing an information recommendation program. When executed by a processor, the information recommendation program implements computer-readable program instructions for the information recommendation method described below, thus solving the technical problem that the passive nature of multimedia information recommendation in related technologies leads to poor recommendation results. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by this invention are the same as those of the information recommendation method provided in the above embodiments, and will not be repeated here.

[0208] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0209] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0210] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0211] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. An information recommendation method, characterized in that, The information recommendation method includes the following steps: Acquire customer segmentation information corresponding to multimedia information and user communication data of at least one information recipient; From among the aforementioned information recipients, determine the target information recipients whose user communication data matches the customer group screening information; At least one target area is determined based on the user communication location of each target information recipient, and the user group preference type score of each target information recipient is determined based on the customer group screening information and the user communication data of each target information recipient. A target user heat map is generated based on each of the areas to be targeted and the user group preference type scores of each of the areas to be targeted. The target user heat map is then sent to the information recommender so that the information recommender can make offline information recommendations based on the target user heat map. The user communication data includes internet communication logs, and the step of determining the user group preference type score for each of the target information recipients based on the customer group screening information and the user communication data of each target information recipient includes: Based on the internet communication logs, determine the user behavior preferences and information conversion channel scores of each of the target information recipients; The user behavior preference score for each of the target information recipients is determined based on the user behavior preferences of each target information recipient and the information conversion channel score. The user group preference type score for each of the regions to be targeted is determined based on the user behavior preference score corresponding to each region. The step of determining the user behavior preference score corresponding to each user behavior preference of each target information recipient based on the user behavior preferences of each target information recipient and the information conversion channel score includes: The user behavior preferences and information conversion channel scores of each target information recipient are input into a preset user behavior scenario preference algorithm to obtain the user behavior preference scores corresponding to each user behavior preference of each target information recipient. The user behavior scenario preference algorithm is as follows: ； wherein p i,j is a user behavior preference score of a user behavior preference j of the information recipient i, x i,k is an information conversion channel score of an information conversion channel k of the information recipient i, w i,j is a preset weight of the user behavior preference j of the information recipient i.

2. The information recommendation method as described in claim 1, characterized in that, The step of determining the target information recipients from among the information recipients whose user communication data matches the customer group screening information includes: From the various information recipients, a preliminary selection of information recipients is determined to match the user communication data with the customer group screening information; Check whether each of the aforementioned initial selection information recipients has a historical scanning record; The initial information recipients with historical scanning records are identified as the target information recipients.

3. The information recommendation method as described in claim 2, characterized in that, Before the step of determining at least one delivery area based on the user communication location of each of the target information recipients, the method further includes: The QR code location information in the historical scanning records corresponding to each of the target information recipients is determined as the user communication location corresponding to each of the target information recipients.

4. The information recommendation method as described in claim 1, characterized in that, After the step of sending the target user heat map to the information recommender, the method further includes: Obtain feedback information from each of the target information recipients; The weights of each user behavior preference of each target information recipient are iteratively updated based on the feedback information.

5. The information recommendation method as described in claim 4, characterized in that, The feedback information includes feedback behavior preferences, and the step of iteratively updating the weights of each user behavior preference of each target information recipient based on the feedback information includes: Based on the correlation between preset behavioral preferences, determine the correlation between the feedback behavioral preferences of each target information receiver and their corresponding user behavioral preferences; The weights of user behavior preferences corresponding to negative correlations are adjusted to be reduced, while the weights of user behavior preferences corresponding to positive correlations are adjusted to be increased.

6. The information recommendation method as described in claim 1, characterized in that, The customer group filtering information includes at least one of information content tags, customer profile tags, and customer location information; The user communication data includes at least one of internet communication logs, user basic attribute data, and user communication location. The step of determining the target information recipients from among the information recipients whose user communication data matches the customer group screening information includes: Determine the user behavior preferences of each information recipient based on the internet communication logs; At least one target information recipient that meets the customer group matching conditions is determined from the information recipients, wherein the customer group matching conditions include at least one of the following: the user behavior preferences match the information content tags, the user basic attribute data matches the customer group profile tags, and the user communication location matches the customer group location information.

7. The information recommendation method as described in claim 6, characterized in that, The steps to obtain information content tags include: Obtain multimedia information, and retrieve the associated information corresponding to the multimedia information through web crawling; The information content tags of the multimedia information are determined based on the associated information.

8. The information recommendation method as described in claim 7, characterized in that, The step of determining the information content tag of the multimedia information based on the association information includes: The associated information is matched with a preset keyword database to determine at least one target keyword corresponding to the multimedia information; Based on the preset mapping relationship between keywords and category tags, determine the target category tag corresponding to each of the target keywords; The number of times the target category label is determined is counted, and the target category label whose number of determinations meets the preset determination number condition is determined as the information content label of the multimedia information.

9. The information recommendation method as described in claim 1, characterized in that, After the step of determining the target information recipient from each of the information recipients that matches the user communication data with the customer group screening information, the method further includes: Information recommendations are made to the target information recipients.

10. The information recommendation method as described in claim 9, characterized in that, The step of determining the target information recipients from among the information recipients whose user communication data matches the customer group screening information includes: From the various information recipients, a preliminary selection of information recipients is determined to match the user communication data with the customer group screening information; Obtain the current delivery frequency of each of the initially selected information recipients; From the selected information recipients, determine the target information recipients whose current delivery frequency meets the preset frequency conditions.

11. An information recommendation device, characterized in that, The information recommendation device includes: a memory, a processor, and an information recommendation program stored in the memory and executable on the processor, wherein the information recommendation program, when executed by the processor, implements the steps of the information recommendation method as described in any one of claims 1 to 10.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an information recommendation program, which, when executed by a processor, implements the steps of the information recommendation method as described in any one of claims 1 to 10.

Images