Information pushing method, electronic device and computer program product

Abstract

This disclosure provides an information push method, an electronic device, and a computer program product. The information push method of this disclosure includes: calculating the feature value of a target property with respect to a first feature; determining the predicted transaction date of the target property; the first feature being a property feature whose influence on the property's transaction rate exceeds a certain threshold; determining target guidance information about the target property and the recipients of the target guidance information based on an information push strategy and the predicted transaction date; and pushing the target guidance information to the recipients.

Description

Technical Field

[0001] This disclosure relates to the technical fields of data processing, and in particular to an information push method and an electronic device computer program product. Background Technology

[0002] Second-hand housing transactions are characterized by long cycles, complex procedures, and numerous influencing factors (such as viewing activity and market fluctuations). Currently, the estimation of transaction time for properties mainly relies on human experience. However, human experience is subject to lag and subjectivity, making it impossible to accurately predict the specific transaction point of a property. This results in properties being in a "black box" state during the listing period, lacking scientific management of the circulation rhythm.

[0003] Existing technologies typically intervene passively only during the contract signing stage after a property transaction is completed, or by pushing indiscriminate marketing information during the listing period. They cannot provide differentiated advice based on the actual speed of the property transaction, which affects the efficiency of the property transaction and may even affect the stability of the transaction. Summary of the Invention

[0004] This disclosure provides an information push method, an electronic device, and a computer program product.

[0005] According to one aspect of this disclosure, an information push method is provided, comprising: calculating a feature value of a target property with respect to a first feature; determining a predicted transaction date for the target property, wherein the first feature is a property feature whose influence on the transaction rate exceeds a certain threshold; determining target guidance information about the target property and the target guidance information to be pushed to the target property based on an information push strategy and the predicted transaction date; and pushing the target guidance information to the target property.

[0006] Based on a technical solution, the transaction date is predicted by combining high-impact characteristics, and a push strategy is dynamically matched to achieve accurate guidance of the transaction process, effectively improving transaction efficiency and user experience.

[0007] In some implementations, based on an information push strategy and the predicted transaction date, determining target guidance information for the target property and the recipients of the target guidance information includes: determining whether the listing duration between the listing date of the target property and the predicted transaction date is greater than or equal to a first duration threshold; and based on the information push strategy, when the listing duration is less than the first duration threshold, generating first target guidance information and determining recipients applicable to the first target guidance information, wherein the first target guidance information is transaction preparation information to ensure the target property is transacted on the predicted transaction date; or, when the listing duration is greater than or equal to the first duration threshold, generating second target guidance information and determining recipients applicable to the second target guidance information, wherein the second target guidance information is a behavioral suggestion to shorten the listing duration.

[0008] According to one technical solution, by distinguishing the relationship between the listing duration and the corresponding threshold, transaction preparation information to ensure the smooth completion of the transaction or behavioral suggestions to accelerate the transaction can be dynamically generated and accurately matched to the target audience, thereby balancing transaction certainty and efficiency and improving the overall conversion effect.

[0009] In some implementations, generating first target guidance information and determining the target recipients for the first target guidance information includes: determining whether the remaining waiting time between the current date and the predicted transaction date is greater than or equal to a second time threshold, wherein the second time threshold is less than the first time threshold; and based on an information push strategy, when the remaining waiting time is greater than or equal to the second time threshold, using property-related information as the first target guidance information and determining the property owner of the target property as the target recipient of the first target guidance information; or, when the remaining waiting time is less than the second time threshold, using transaction preparation information as the first target guidance information and determining the business personnel of the target property as the target recipient of the first target guidance information.

[0010] According to one technical solution, the content and recipients of the first target guidance information are dynamically refined based on the remaining waiting time. In the early stages of the transaction, property-related information is pushed to the property owner to maintain attention, and transaction preparation information is pushed to the business personnel as the transaction approaches to ensure performance, thereby achieving phased and precise transaction process coordination.

[0011] In some implementations, after generating the second target guidance information and determining the target recipients for the second target guidance information, the method further includes: using property optimization suggestions and financial service information applicable to the target property as the second target guidance information, determining the property owner of the target property as the target recipient of the second target guidance information; and using operational behavior suggestions for the target property as another second target guidance information, determining the business personnel as the target recipient of the other second target guidance information.

[0012] According to one aspect of the technical solution, after generating the second target guidance information, the system pushes property optimization suggestions and financial service information to property owners, while pushing operational behavior suggestions to business personnel, thereby achieving differentiated intervention for different roles and collaboratively shortening the listing time and accelerating the transaction process.

[0013] In some implementations, before calculating the feature value of the target property with respect to the first feature and determining the predicted transaction date of the target property, the method includes: determining at least one first feature among a plurality of property features, the first feature being the number of viewings or the number of interviews, wherein the number of viewings is the total number of visits to the property between the listing date and the predicted transaction date of the property; and the number of interviews is the total number of offline conversations between the property owner and the viewer between the listing date and the predicted transaction date of the property.

[0014] According to one technical solution, by using key behavioral data reflecting market activity, such as the number of viewings and face-to-face meetings, as the primary feature, it is possible to more accurately characterize the transaction rate of properties and improve the accuracy of predicting transaction dates.

[0015] In some implementations, determining at least one first feature among multiple property features includes: using an impact analysis model to analyze each property feature of sold properties in historical transaction records, determining the degree of influence of each property feature on the property transaction rate, wherein the property features include static features and dynamic features, the static features characterizing the physical attributes of the property, and the dynamic features including viewing records and interview records, wherein the viewing records record the number of people viewing the property per day and the corresponding viewing dates from the listing date of the sold property to the predicted transaction date of the sold property, and the interview records record the number of offline conversations per day and the corresponding conversation dates from the listing date of the sold property to the predicted transaction date of the sold property; and when a property feature whose influence exceeds the degree threshold is a dynamic feature, calculating the quantitative result corresponding to the dynamic feature and using the quantitative result as the first feature; and when a property feature whose influence exceeds the degree threshold is a static feature, using the static feature as the first feature.

[0016] Based on one technical solution, by introducing an impact degree analysis model, the impact of static and dynamic features on the transaction rate is systematically evaluated, the key primary features are accurately selected, and unstructured behavioral data such as viewing records and interview records are quantified. This achieves an effective transformation from raw behavioral logs to modelable features, significantly improving the scientific nature of feature selection and the accuracy of model prediction.

[0017] In some implementations, when the property feature whose influence exceeds the threshold is a dynamic feature, the quantitative results corresponding to the dynamic feature are statistically analyzed, including: when the property feature whose influence exceeds the threshold is a viewing record, calculating the sum of the number of viewings per day from the listing date of the sold property to the predicted transaction date of the sold property, to obtain the total number of viewings, the total number of viewings being the quantitative result of the viewing record; and / or when the property feature whose influence exceeds the threshold is a face-to-face meeting record, calculating the sum of the number of offline conversations per day from the listing date of the sold property to the predicted transaction date of the sold property, to obtain the total number of offline conversations, the total number of offline conversations being the quantitative result of the face-to-face meeting record.

[0018] According to one technical solution, by accumulating and summing dynamic behavioral data such as viewing records and interview records, unstructured time-series interaction information is transformed into computable numerical features, effectively capturing the activity of the housing market and improving the accuracy and interpretability of the model's prediction of transaction rate.

[0019] In some implementations, calculating the feature value of the target property with respect to a first feature and determining the predicted transaction date of the target property includes: calling a duration analysis model to analyze the property features including the feature value of the first feature, generating a listing duration for the target property; and superimposing the listing date of the target property with the listing duration to determine the predicted transaction date.

[0020] Based on one technical solution, the key characteristics of the target property are comprehensively analyzed by calling the duration analysis model, the listing duration is accurately predicted, and the predicted transaction date is calculated by combining the listing date, thus realizing dynamic prediction of the transaction node.

[0021] In some implementations, before calling the duration analysis model to analyze the property features including the feature values ​​of the first feature, the process includes: determining multiple first features for each sold property in historical transaction records, wherein the first feature is the number of viewings or interviews; and training the original analysis model using the first features of the sold properties and the actual listing duration of the corresponding sold properties to obtain the duration analysis model, wherein the duration analysis model can predict the listing duration of the corresponding property based on the feature values ​​of the first feature.

[0022] According to one technical solution, by training the original analysis model based on the correlation between key features (such as the number of viewings and face-to-face meetings) and the actual listing duration in historical transaction data, a duration analysis model that can accurately predict the listing duration of target properties is constructed, providing reliable model support for subsequent prediction of transaction dates and accurate delivery of target guidance information.

[0023] According to one technical solution, by combining the offset duration mechanism in the information push strategy, a key time window is set before the transaction date. The information push is only triggered when the candidate push date falls after the listing date, which ensures the rationality and effectiveness of the service intervention timing, avoids the waste of resources caused by pushing too early or too late, and improves the reach and conversion efficiency of financial service information.

[0024] According to another aspect of this disclosure, an electronic device is provided, comprising: a memory storing execution instructions; and a processor executing the execution instructions stored in the memory, causing the processor to perform the information push method according to any embodiment of this disclosure.

[0025] According to another aspect of this disclosure, a readable storage medium is provided, wherein executable instructions are stored therein, which, when executed by a processor, are used to implement the information push method described in any embodiment of this disclosure.

[0026] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the information push method described in any embodiment of this disclosure. Attached Figure Description

[0027] The accompanying drawings illustrate exemplary embodiments of the present disclosure and, together with the description thereof, serve to explain the principles of the present disclosure. These drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification.

[0028] Figure 1 This is a schematic diagram illustrating an application scenario of the information push method according to the embodiments of this disclosure.

[0029] Figure 2 This is a flowchart of an information push method according to an embodiment of the present disclosure.

[0030] Figure 3 This is a flowchart of multi-terminal interaction during information push according to the embodiments of this disclosure.

[0031] Figure 4 This is a flowchart illustrating the process of determining the first feature according to an embodiment of the present disclosure.

[0032] Figure 5 This is a flowchart illustrating the generation process of the duration analysis model according to the embodiments of this disclosure.

[0033] Figure 6 This is a schematic diagram of various date relationships according to embodiments of this disclosure.

[0034] Figure 7 This is a schematic block diagram of the structure of an information push device according to an embodiment of the present disclosure.

[0035] Figure 8 This is a schematic block diagram of an electronic device according to one embodiment of the present disclosure. Detailed Implementation

[0036] The present disclosure will now be described in further detail with reference to the accompanying drawings and examples. It should be understood that the specific examples described herein are for illustrative purposes only and are not intended to limit the scope of the disclosure. Furthermore, it should be noted that, for ease of description, only the parts relevant to the present disclosure are shown in the accompanying drawings.

[0037] It should be noted that, where there is no conflict, the embodiments and features described in this disclosure can be combined with each other. The technical solutions of this disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0038] Second-hand housing transactions are characterized by long cycles, numerous stages, and complex influencing factors. The transaction process is not only affected by the property's own attributes (such as price, layout, and decoration), but also highly dependent on external dynamic factors, such as viewing activity, regional market fluctuations, and the psychological expectations of both buyers and sellers. Currently, the industry's judgment on the timing of property transactions mainly relies on the personal experience of sales staff or property owners. This method generally suffers from problems such as lag, strong subjectivity, and poor replicability, making it difficult to accurately identify key transaction nodes. As a result, many properties remain in a "black box" state during the listing period—it is impossible to predict when a transaction might occur, and there is a lack of proactive management mechanisms based on the transaction rhythm.

[0039] Existing technical solutions are mostly limited to two extremes: one is passively intervening only when the transaction is nearing completion and entering the signing stage, missing the early guidance window; the other is conducting indiscriminate and generalized information pushes (such as uniformly sending marketing content or general service introductions) from the initial listing stage, failing to differentiate and dynamically respond based on the actual transaction speed of the property. This "one-size-fits-all" strategy cannot effectively stimulate market attention and optimize property performance when transaction expectations are too long, nor can it coordinate all parties to prepare for performance in a timely manner when the transaction is nearing completion. This not only reduces overall transaction efficiency but may also affect transaction stability due to untimely response or improper intervention.

[0040] Therefore, this disclosure proposes an information push method.

[0041] Figure 1 This is a schematic diagram illustrating an application scenario of the information push method according to the embodiments of this disclosure. For example... Figure 1 As shown, this application scenario may include a server 100 and a terminal device 200. The server 100 and the terminal device 200 can interact with each other via a network connection. The server 100 can be a cloud server or a physical server, and the terminal device 200 can be a smart device such as a computer, mobile phone, or tablet. The server 100 can receive requests from the terminal device 200 and run the main information push method disclosed herein.

[0042] Figure 2 This is a flowchart of an information push method according to an embodiment of this disclosure. For example... Figure 2 As shown, this disclosure proposes an information push method M2OO, which accurately predicts the transaction date by focusing on high-impact characteristics, and dynamically matches the target guidance information and push recipients accordingly, thereby achieving proactive and differentiated intervention in the transaction process and effectively improving transaction efficiency and stability.

[0043] Step S202: Calculate the feature value of the target property with respect to the first feature, and determine the predicted transaction date of the target property.

[0044] Target properties refer to properties that are for sale and require mortgage release. Typically, a property becomes a target property as soon as it is registered and listed for sale on a real estate sales platform, used for subsequent prediction and analysis of the transaction date. During property registration, static information describing its physical attributes must be provided, such as building area, unit price, property use (e.g., owner-occupied, rented, or vacant), and property type (e.g., residential, apartment, commercial). During the sales process, if potential buyers express interest or inquire, platform staff will arrange viewings or meetings with the property owner. Each viewing and meeting must be accurately recorded by staff, creating separate viewing and meeting records. These records are dynamically updated based on actual interactions, serving as an important dynamic characteristic reflecting the property's level of interest and effectively characterizing its market activity.

[0045] The primary feature refers to property characteristics that significantly impact the transaction rate beyond a preset threshold. There can be one or more primary features. When a feature exceeding this threshold is a dynamic feature (such as viewing records or interview records), it needs to be quantified, and the result used as the primary feature. Since dynamic features typically exist as behavioral records, they cannot be directly used for model calculations and must be converted into numerical indicators through statistical methods and aggregation. Static features (such as area, unit price, and property rights) can be directly used as primary features because they possess a directly usable structured data format. The threshold can be flexibly set according to actual business needs, and multiple key features can be selected, collectively forming the core feature set affecting the transaction rate.

[0046] The process of quantifying dynamic characteristics is as follows: When the property characteristic with an impact exceeding the threshold is a viewing record, the total number of viewings per day is calculated from the listing date of the sold property to the predicted transaction date of the sold property, which is the quantification result of the viewing record; and / or when the property characteristic with an impact exceeding the threshold is a face-to-face meeting record, the total number of offline conversations per day is calculated from the listing date of the sold property to the predicted transaction date of the sold property, which is the quantification result of the face-to-face meeting record.

[0047] Eigenvalues ​​refer to the values ​​of the primary feature for a specific target property, and these eigenvalues ​​vary from property to property. For example, when "area" is determined as the primary feature, the eigenvalue for target property A is 100 square meters, while the eigenvalue for property B is 140 square meters. Eigenvalues ​​reflect the specific performance of a property on a particular key feature and are important inputs for subsequent decision-making regarding transaction cycle prediction and information delivery.

[0048] Step S204: Based on the information push strategy and the predicted transaction date, determine the target guidance information for the target property and the recipients of the target guidance information.

[0049] The information push strategy includes the relationship between the listing duration of the target property and the target guidance information and the target audience, as well as the relationship between the remaining waiting time of the target property and the target guidance information and the target audience. For example, if the listing duration of a target property exceeds the first duration threshold (e.g., 60 days), it can be determined that its transaction pace is slow, triggering the "accelerated facilitation" strategy. The strategy should push property optimization suggestions to the property owner, such as adding a detailed description of the property, improving property photos, or suggesting interior design improvements. Operational suggestions should also be pushed to sales personnel, such as increasing viewing arrangements and enhancing exposure. If the listing duration is less than the first duration threshold, but the remaining waiting time (i.e., from the current date to the predicted transaction date) is less than the second duration threshold (e.g., 7 days), the strategy switches to "performance guarantee," pushing transaction preparation information (such as a list of required materials and reminders about the transfer process) to sales personnel to ensure the successful transaction of the target property.

[0050] Information push strategies can be based on historical data analysis, such as tracing the transaction timeline of past properties and behavioral data like click-through rates and inquiry conversion rates of targeted information; they can also be achieved through questionnaires, user interviews, and other methods to collect a large amount of data on property owners' service perceptions, changing needs, and decision-making psychology at each stage of the transaction. Furthermore, information push strategies can be dynamically adjusted to suit different cities, property types, or loan-to-value ratios to improve the accuracy and adaptability of the push notifications.

[0051] The listing date refers to the date the target property is registered for sale on the real estate sales platform, typically expressed as year, month, and day. The predicted transaction date also indicates the year, month, and day of the expected transaction for the target property based on the first feature. The push date represents the peak demand for financial services information from the property owner between the listing date and the predicted transaction date, also expressed as year, month, and day.

[0052] The target guidance information includes primary target guidance information and secondary target guidance information. The primary target guidance information is transaction preparation information to ensure that the target property is sold on the predicted transaction date. The secondary target guidance information is behavioral suggestions to shorten the listing period. By distinguishing between the two types of guidance logic, namely "ensuring a sale" and "facilitating a sale," dynamic and precise intervention can be achieved throughout the entire lifecycle of the property, balancing transaction efficiency and stability.

[0053] The primary objective is to guide information to ensure contract fulfillment. As the predicted transaction date approaches, information such as a list of signing documents, key points for property rights verification, and guidelines for the transfer process should be provided to ensure efficient collaboration among all parties and a smooth transaction.

[0054] The second objective focuses on accelerating the conversion process. When the listing period is too long and the transaction expectation is lagging behind, targeted behavioral suggestions are pushed, such as optimizing the property display (adding detailed descriptions of the property by the property owner, updating pictures, adding soft furnishings), improving exposure strategies, or strengthening viewing arrangements, in order to activate market response and shorten the actual transaction cycle.

[0055] The target audience for the push notifications is dynamically determined based on the content of the guidance information, and typically includes the property owner or relevant sales personnel of the target property. The process of matching the target audience with the guidance information will be described later and will not be elaborated here.

[0056] Step S206: Push target guidance information to the target object.

[0057] Specifically, targeted guidance information can be pushed to the target audience through multiple channels (such as SMS, APP messages, telephone, or dedicated account manager contact).

[0058] Figure 3 This is a flowchart illustrating the multi-terminal interaction process during information push according to the embodiments of this disclosure. For example... Figure 3 As shown, the implementation of the information push method involves collaboration and data interaction among multiple processing modules, including the coordinated operation of core algorithm models such as the impact analysis model and the transaction duration prediction model. Simultaneously, historical transaction data needs to be retrieved from the database for model input and analysis. After predicting the transaction date and determining the push timing, the generation of target guidance information is further triggered, providing customized content to the push recipients through appropriate channels, including the mobile phones, computers, and other terminal devices of property owners. The entire process achieves a closed-loop linkage of data collection, intelligent analysis, and service delivery, ensuring the accuracy and timeliness of information push.

[0059] Specifically, in step 301, historical transaction records are first retrieved from the database and input into the influence analysis model of various characteristics of the sold properties. Then, in step 302, the influence analysis model identifies the first characteristic that significantly impacts the transaction rate. In step 303, all or a large number of historical transaction records are retrieved from the database, and the first characteristic and actual listing duration of each sold property are determined from these records. This data is used to train the original analysis model to obtain a duration analysis model. In step 304, the first characteristic of the target property is input into the duration analysis model. In step 305, the duration of listing for the target property is estimated through analysis of the first characteristic using the duration analysis model. Finally, in step 306, the predicted transaction date is determined based on the predicted listing duration, and the appropriate target guidance information and its target audience are dynamically determined based on the predicted transaction date. This upgrades the system from "experience-based judgment" to "intelligent prediction," effectively improving the certainty, efficiency, and coordination of second-hand housing transactions.

[0060] In some implementations, a duration analysis model is invoked to analyze the property features, including the feature values ​​of the first feature, to generate the listing duration for the target property; and the listing date of the target property is overlaid with the listing duration to determine the predicted transaction date.

[0061] In some implementations, the process of determining the target guidance information for the target property and the recipients of the target guidance information involves: determining whether the listing duration of the target property from its listing date to the predicted transaction date is greater than or equal to a first duration threshold; and based on an information push strategy, generating first target guidance information when the listing duration is less than the first duration threshold, and determining the recipients of the first target guidance information, wherein the first target guidance information is transaction preparation information to ensure that the target property is transacted on the predicted transaction date; or, generating second target guidance information when the listing duration is greater than or equal to the first duration threshold, and determining the recipients of the second target guidance information, wherein the second target guidance information is a behavioral suggestion to shorten the listing duration.

[0062] Specifically, when the listing duration is less than a first duration threshold, it is determined whether the remaining waiting time between the current date and the predicted transaction date is greater than or equal to a second duration threshold, where the second duration threshold is less than the first duration threshold; and based on the information push strategy, when the remaining waiting time is greater than or equal to the second duration threshold, property-related information is used as the first target guidance information, and the property owner of the target property is identified as the recipient of the first target guidance information; or, when the remaining waiting time is less than the second duration threshold, transaction preparation information is used as the first target guidance information, and the business personnel of the target property are identified as the recipients of the first target guidance information.

[0063] For example, let's set the first time threshold at 60 days and the second time threshold at 15 days. The target property was listed on November 1, 2025, with a predicted listing period of 50 days and a predicted transaction date of December 20, 2025. Clearly, the listing period for this target property is less than the first time threshold, indicating that the property is in a normal transaction cycle. In this case, it's necessary to further determine the remaining waiting time. If the current date is December 1, 2025, the remaining waiting time is 19 days, which is greater than the second time threshold (15 days). "Transaction preparation information" (such as regional weekly transaction reports, price trends for similar unit types, etc.) will be used as the first target guidance information and pushed to the property owner to help them continuously monitor market dynamics. If the current date is December 10, 2025, the remaining waiting time is 10 days. Then, "transaction preparation information" (such as a list of signing materials, transfer process guidelines, etc.) will be used as the first target guidance information and pushed to the sales staff to ensure a smooth transaction.

[0064] When a target property requires mortgage release, transaction preparation information can also include financial service information. This financial service information refers to tailored financial solutions and related services for the target property, designed to help the property owner smoothly complete the mortgage release and transaction process. This type of information typically includes: introductions to mortgage release loan products, loan amounts and interest rate plans, fund disbursement timelines, application requirements and required documents, repayment arrangements, descriptions of financial institution qualifications, and personalized financial service recommendations based on the property owner's creditworthiness and property valuation. Furthermore, it can cover supporting financial services throughout the transaction process, such as fund supervision, tax payment advances, and mortgage loan pre-approval. By providing comprehensive, accurate, and actionable financial information, it helps property owners obtain timely financial support at key points in the transaction, improving transaction efficiency and service experience.

[0065] When the listing duration is greater than or equal to the first duration threshold, it also includes: using property optimization suggestions and financial service information applicable to the target property as the second target guidance information, and identifying the property owner of the target property as the recipient of the second target guidance information; and using operational behavior suggestions for the target property as another second target guidance information, and identifying the business personnel as the recipient of the other second target guidance information.

[0066] For example, if the listing date of the target property is September 1, 2025, and the predicted transaction date is December 10, 2025, its listing period is as long as 100 days, which obviously exceeds the first duration threshold. This triggers the generation of the second target guidance information, namely, pushing property optimization suggestions and supporting service information to the property owner, and pushing operational behavior suggestions (such as increasing viewings and increasing exposure) to the business personnel.

[0067] Similarly, the behavioral recommendations are not limited to those mentioned above; the preceding text is merely one feasible implementation method.

[0068] This solution employs a tiered judgment mechanism to refine guidance strategies based on the remaining waiting time, avoiding logical redundancy and information mismatch. It can proactively intervene and accelerate sales when transaction expectations are too long, and accurately push preparation information and ensure performance as the transaction approaches, achieving efficient resource allocation and intelligent collaboration throughout the entire transaction process.

[0069] In some implementations, when the target guidance information is financial service information, a push date can be determined between the listing date and the predicted transaction date, and the target guidance information can be pushed to the property owner on the push date. In this case, the information push strategy also records the correlation between the peak demand for financial service information and the predicted transaction date, where the push date is the date when the property owner's demand for financial service information reaches its peak.

[0070] Based on the information push strategy, the push date for financial service information is determined between the listing date of the target property and the predicted transaction date, including: when the information push strategy is to use a date earlier than the predicted transaction date by an offset period as the push date, a date earlier than the transaction date of the target property by an offset period is determined as a candidate date; and when the candidate date is later than the listing date of the target property, the candidate date is used as the push date for financial service information.

[0071] Figure 4 This is a flowchart illustrating the process of determining the first feature according to an embodiment of this disclosure. The following is in conjunction with... Figure 4 This paper introduces the impact analysis model mentioned above and explains its processing logic.

[0072] The impact analysis model, based on a large amount of historical transaction data in a database, can quantify the impact of various property characteristics on the transaction rate, and select key factors with significant influence as the primary feature. "Significant impact" refers to a feature whose influence on the transaction speed exceeds a preset threshold. This model can be implemented using survival analysis models, such as the Cox proportional hazards model, which models the time span from listing to transaction (i.e., "survival time") to assess the strength of the impact of various static and dynamic features on the probability of transaction at different time points. This model can capture the dynamic effects of features changing over time, thus more accurately identifying the core factors that truly drive the transaction rate.

[0073] Specifically, in step 401, the influence analysis model is invoked to analyze the characteristics of each property in the historical transaction records to determine the degree of influence of each characteristic on the transaction rate. The data used can be several second-hand housing transaction samples from the previous signing cycle (the signing cycle can be a calendar year), covering the following static characteristics: region, house area, house use (e.g., owner-occupied / rental / vacant), and house ownership type (e.g., residential / apartment / commercial). It also includes time characteristics: listing date (first listing date), actual transaction date, and offset duration; where the offset duration is the length of time between the actual transaction date and the date of push notification of financial service information. It also includes dynamic characteristics: viewing records and interview records. Viewing records include viewing date and number of viewings per day; interview records include interview date and number of interviews per day.

[0074] The analysis results from the influence degree analysis model show that: for every unit increase in the number of viewings, the transaction rate increases by more than 1 times, indicating that higher attention from prospective buyers significantly accelerates the transaction process; for every unit increase in the number of face-to-face meetings, the transaction rate increases by more than 2 times, indicating that in-depth communication has a stronger positive driving effect on facilitating transactions.

[0075] In step 402, it is determined whether the influence of the currently analyzed property feature exceeds a preset threshold. If the influence exceeds the threshold and the feature is a dynamic feature (such as viewing records or interview records), then step 403 is executed to confirm its dynamic attributes, and then step 404 is performed to quantify it: for example, the original viewing records are summarized by time interval to the total number of viewings, and the interview records are summarized to the total number of offline conversations, thus converting them into numerical indicators. Subsequently, in step 405, this quantification result is used as the first feature.

[0076] If the feature whose influence exceeds the threshold is a static feature (such as area, property rights, etc.), it is directly used as the first feature in step 406, because it already has a structured form that can be directly input into the model.

[0077] In some implementations, the first feature can be the number of viewings or the number of face-to-face meetings, where: the number of viewings is defined as the total number of visits from the listing date to the predicted transaction date; the number of face-to-face meetings is defined as the total number of offline conversations between the property owner and potential buyers within the same time period. These quantified core dynamic features, combined with static attributes, constitute the key inputs to the prediction model, significantly improving the accuracy of predicting the transaction rate of properties.

[0078] Figure 5 This is a flowchart illustrating the generation process of the duration analysis model according to the embodiments of this disclosure. The following is in conjunction with... Figure 5 This paper introduces the duration analysis model and explains the processing logic involved.

[0079] In step 501, the properties that have already been sold are cleaned.

[0080] First, the historical transaction records in the database are cleaned. This step aims to remove logically contradictory samples, such as properties where the actual transaction date is earlier than the property certificate date, the actual transaction date is earlier than the listing date, or the listing date is earlier than the property certificate date. Simultaneously, missing values ​​are filled: numeric fields are filled with the mean, and categorical fields are filled with the mode.

[0081] In step 502, the remaining sold properties are formatted.

[0082] After data cleaning, the remaining sold properties are formatted, including core indicator calculations (such as defining listing duration as the time difference between the actual transaction date and the listing date) and feature engineering (such as category feature coding and combined feature construction). This stage ensures data consistency and structure, providing high-quality data input for subsequent training of the original analysis model.

[0083] In step 503, the sold properties are divided into a training set and a validation set.

[0084] The formatted data is randomly divided into a training set and a validation set in an 8:2 ratio to maintain consistency in the distribution of listing duration and region between the two sets, thus avoiding sampling bias. The training set is used for model parameter learning, and the validation set is used to evaluate the model's generalization ability. Of course, the ratio of the training set to the validation set can be set according to needs; the above is just an example.

[0085] In step 504, the training set is used to train the original analysis model.

[0086] The original analytical model is trained using the training set. The model's input includes multiple selected primary features and their corresponding feature values, such as region, house area, house use, property type, building age, number of viewings, number of interviews, and historical transaction prices (one or more of these). These features collectively constitute the key variables affecting the property transaction cycle. The model's output target is the predicted listing duration (i.e., the number of days from listing to actual transaction).

[0087] During training, the model continuously optimizes its internal parameters by learning the mapping relationship between input features and actual transaction cycles. Specifically, the model is first initialized with default parameters, a maximum number of iterations is set to, for example, 1000 iterations, and MSE (Mean Squared Error) is selected as the objective function. MSE is one of the most commonly used evaluation metrics in regression models, calculated as the average of the squared differences between the predicted and actual values. The smaller the MSE, the closer the model's prediction is to the actual value, and the higher the prediction accuracy.

[0088] During training, each iteration calculates the MSE of the current model on the training set and uses this to backpropagate the error through optimization algorithms such as gradient descent, gradually adjusting the model's weights and parameters to minimize the loss function. Simultaneously, to prevent overfitting (i.e., performing well on the training set but generalizing poorly on new data), an early stopping mechanism is introduced; when the MSE of the model on the validation set no longer decreases within several consecutive iterations (e.g., 50 iterations), the training process automatically terminates, preserving the optimal parameter state.

[0089] While initial training can achieve optimal model performance, further hyperparameter tuning (such as learning rate, tree depth, and subsampling rate) using a validation set is necessary to further improve prediction accuracy and explore better model configurations. Therefore, MSE (Mean Sequence Estimation) is not only an optimization goal during training but also a core basis for evaluating model performance and guiding hyperparameter tuning, permeating the entire model development process to ensure that the final model has good fitting ability and generalization performance.

[0090] In step 505, it is determined whether the model's MSE on the training set is greater than the error threshold.

[0091] If the error exceeds the error threshold, proceed to the next step of adjusting the model's key parameters; if the error is less than or equal to the error threshold, directly evaluate the model's performance using the validation set. The error threshold can be adjusted according to actual needs.

[0092] In step 506, the key parameters of the model are adjusted.

[0093] Key parameters of the model, including tree depth, learning rate, minimum number of samples per leaf node, and feature subsampling rate, were tuned using a combination of grid search and Bayesian optimization. The aim was to minimize the MSE (Minimum Selective Equation) on the validation set and improve the model's prediction accuracy.

[0094] In step 507, the model's performance is evaluated using the validation set. The optimized model is used to predict features on the validation set, outputting the predicted listing duration for each property. Evaluation metrics such as MSE and R² are calculated to measure the model's predictive accuracy and fit. R² is the proportion of listing duration variability that the model can explain. In other words, it measures the extent to which the model's predictions "reproduce" or "explain" the actual data's trend. R² values ​​range from 0 to 1 (and may be negative in special cases, indicating extremely poor model performance). An R² of 1 indicates that the model perfectly fits the data, and all predicted values ​​are completely consistent with the actual values. An R² of 0 indicates that the model's predictive ability is equivalent to directly predicting using the average of the target variable, meaning the model provides no additional explanatory power.

[0095] In step 508, it is determined that the evaluation indicators are within the expected range.

[0096] Determine if the evaluation metrics meet business requirements (e.g., MSE ≤ 120 and R² > 0.8). If they do, obtain and apply the duration analysis model; otherwise, return to step 506 to readjust the features or parameters.

[0097] In step 509, the duration analysis model is obtained and applied.

[0098] Once the trained and evaluated model is deployed to a real estate sales platform, it supports inputs such as basic attributes, time information, core behavioral data, and price information for individual second-hand properties, and outputs the predicted transaction time for the property.

[0099] In step 510, the application status is monitored.

[0100] After the model is deployed to the production environment, its application status is continuously monitored, including the accuracy of prediction results and the stability of system operation, to ensure that the model can stably and efficiently serve real business scenarios.

[0101] In addition, the statistical distribution characteristics of the model's prediction results should be monitored in real time, such as the mean, variance, and quantiles (e.g., median, 90th percentile) of the predicted transaction duration. By setting a reasonable fluctuation threshold (e.g., mean ± 3 standard deviations), it is possible to identify whether there are significant shifts or abnormal fluctuations in the prediction distribution, promptly detect potential data drift or model performance degradation issues, and ensure the stability and reliability of the model output.

[0102] Furthermore, we continuously collect real business data such as actual transaction dates of properties, and summarize and analyze the deviation between model predictions and actual results weekly. We calculate key evaluation indicators such as the absolute error (MAE) and mean squared error (MSE) between the predicted listing duration and the actual duration. Based on the analysis results, we generate a model performance evaluation report to comprehensively evaluate the model's prediction accuracy and business adaptability, providing data support and decision-making basis for subsequent model iterations, feature optimization, and parameter tuning.

[0103] Figure 6 This is a schematic diagram illustrating various date relationships according to embodiments of this disclosure. The following is in conjunction with… Figure 6 It provides an intuitive explanation of the relationships between various dates.

[0104] Listing Date A represents the date the property was first listed, marking the beginning of the timeline. Transaction Date B represents the predicted transaction date, the end of the timeline. The total duration from listing Date A to transaction Date B is the listing duration t (indicated by the red curve in the graph), reflecting the property's overall listing period in the market. Offset Duration m represents the pre-set lead time in the information push strategy, indicating the peak demand for targeted guidance information from the target audience (indicated by the green curve in the graph). Push Date C = B - m, meaning the targeted guidance information is provided to the target audience m days before the expected transaction date, ensuring a high degree of alignment between service intervention and the target audience's critical decision-making period. By systematically analyzing the listing duration t, offset duration m, and the degree of alignment between the transaction probability curve and the demand curve, the push strategy can be dynamically optimized, further improving the service's foresight and accuracy. This ensures that targeted guidance information is provided to the target audience at the most appropriate time, thereby enhancing user trust, increasing response conversion rates, and achieving a comprehensive upgrade from passive response to proactive intelligent service.

[0105] Figure 7 This is a schematic block diagram of the structure of an information push device according to an embodiment of this disclosure. For example... Figure 7The information push device 700 shown includes: a date prediction module 710, used to calculate the feature value of the target property with respect to a first feature, and determine the predicted transaction date of the target property, wherein the first feature is a property feature whose influence on the transaction rate exceeds a certain threshold; a second information determination module 720, used to determine the target guidance information and the target guidance information push recipients based on the information push strategy and the predicted transaction date; and a push module 730, used to push the target guidance information to the push recipients.

[0106] The information push device 700 disclosed herein can be in the form of computer software, and each module of the information push device 700 can be in the form of computer software modules.

[0107] The various modules of the information push device 700 disclosed herein are set up to implement the various steps of the information push method. Their execution principles and steps can be referred to the above text and will not be repeated here.

[0108] Figure 8 This is a schematic block diagram of an electronic device according to one embodiment of the present disclosure. Figure 8 As shown, this disclosure also provides an electronic device 1000, including: a processor 1200 and a memory 1300, the memory 1300 storing execution instructions; the processor 1200 executes the execution instructions stored in the memory 1300, causing the processor 1200 to execute an information push method.

[0109] The hardware architecture of the electronic device 1000 can be implemented using a bus architecture. The bus architecture can include any number of interconnect buses and bridges, depending on the specific application of the hardware and overall design constraints. Bus 1100 connects various circuits, including one or more processors 1200, memory 1300, and / or hardware modules. Bus 1100 can also connect various other circuits 1400, such as peripheral devices, voltage regulators, power management circuits, external antennas, etc.

[0110] Bus 1100 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Component (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, only one connection line is used in this diagram, but this does not imply that there is only one bus or only one type of bus.

[0111] This disclosure also provides a readable storage medium storing a computer program that, when executed by a processor, is used to implement the methods described above. A "readable storage medium" can be any means capable of containing, storing, communicating, propagating, or transmitting a program for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples of a readable storage medium include: an electrical connection with one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and portable read-only memory (CDROM), etc.

[0112] This disclosure also provides a computer program product, the methods of which can be implemented wholly or partially through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented wholly or partially as a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed, all or part of the processes or functions of this disclosure are performed.

[0113] Computer programs or instructions can be stored in a readable storage medium or transferred from one readable storage medium to another. For example, the computer program or instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The readable storage medium can be any available medium capable of access, or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; an optical medium, such as a digital video optical disc; or a semiconductor medium, such as a solid-state drive. The computer-readable storage medium can be a volatile or non-volatile storage medium, or it can include both volatile and non-volatile types of storage media.

[0114] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, electronic devices, readable storage media, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0115] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0116] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0117] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0118] In the description of this specification, the references to terms such as "one embodiment / mode," "some embodiments / modes," "example," "specific example," or "some examples," etc., refer to specific features, structures, or characteristics described in connection with that embodiment / mode or example, which are included in at least one embodiment / mode or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment / mode or example. Moreover, the specific features, structures, or characteristics described may be combined in any suitable manner in one or more embodiments / modes or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments / modes or examples described in this specification, as well as the features of different embodiments / modes or examples.

[0119] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0120] Those skilled in the art should understand that the above embodiments are merely for illustrating the present disclosure and are not intended to limit the scope of the disclosure. Those skilled in the art can make other changes or modifications based on the above disclosure, and these changes or modifications still fall within the scope of the present disclosure.

Claims

1. An information push method, characterized in that, include: Calculate the feature value of the target property with respect to the first feature, and determine the predicted transaction date of the target property, wherein the first feature is a property feature whose influence on the transaction rate exceeds a certain threshold; Based on the information push strategy and the predicted transaction date, determine the target guidance information about the target property and the target recipients of the target guidance information; as well as The target guidance information is pushed to the target recipient.

2. The information push method according to claim 1, characterized in that, Based on the information push strategy and the predicted transaction date, the target guidance information for the target property and the recipients of the target guidance information are determined, including: Determine whether the listing duration between the target property's listing date and the predicted transaction date is greater than or equal to a first duration threshold; and Based on the information push strategy, when the listing duration is less than the first duration threshold, a first target guidance information is generated, and the target audience for the first target guidance information is determined. The first target guidance information is transaction preparation information to ensure that the target property is traded on the predicted transaction date. Alternatively, when the listing duration is greater than or equal to the first duration threshold, a second target guidance information is generated, and the target audience for the second target guidance information is determined. The second target guidance information is a behavioral suggestion to shorten the listing duration.

3. The information push method according to claim 2, characterized in that, Generate first target guidance information and determine the push target applicable to the first target guidance information, including: Determine whether the remaining waiting time between the current date and the predicted transaction date is greater than or equal to a second time threshold, wherein the second time threshold is less than the first time threshold; and Based on the information push strategy, when the remaining waiting time is greater than or equal to the second time threshold, the property-related information is used as the first target guidance information, and the property owner of the target property is identified as the target recipient of the first target guidance information; or, when the remaining waiting time is less than the second time threshold, the transaction preparation information is used as the first target guidance information, and the business personnel of the target property are identified as the target recipient of the first target guidance information.

4. The information push method according to claim 2, characterized in that, After generating the second target guidance information and determining the push recipients applicable to the second target guidance information, the process also includes: Using property optimization suggestions and applicable financial service information for the target property as second target guidance information, the property owner of the target property is identified as the recipient of the second target guidance information; and The operational behavior suggestions regarding the target property are used as another second target guidance information, and the business personnel are identified as the recipients of the second target guidance information.

5. The information push method according to claim 1, characterized in that, Before calculating the feature value of the target property with respect to the first feature and determining the predicted transaction date of the target property, the process includes: Among multiple property characteristics, at least one first characteristic is identified, wherein the first characteristic is the number of viewings or the number of interviews, wherein the number of viewings is the total number of people viewing the property between the listing date and the predicted closing date of the property; and the number of interviews is the total number of conversations between the property owner and the viewer between the listing date and the predicted closing date of the property.

6. The information push method according to claim 5, characterized in that, Identify at least one primary feature among multiple property features, including: An impact analysis model is used to analyze the characteristics of each sold property in historical transaction records to determine the degree of influence of each property characteristic on the property's transaction rate. These characteristics include static and dynamic features. Static features represent the physical attributes of the property, while dynamic features include viewing records and interview records. The viewing records document the number of people viewing the property per day and the corresponding viewing dates from the listing date to the predicted transaction date. The interview records document the number of offline conversations per day and the corresponding conversation dates from the listing date to the predicted transaction date. When a property feature whose influence exceeds the threshold is a dynamic feature, the quantitative result corresponding to the dynamic feature is statistically analyzed, and the quantitative result is used as the first feature; and when a property feature whose influence exceeds the threshold is a static feature, the static feature is used as the first feature.

7. The information push method according to claim 6, characterized in that, When a property feature whose impact exceeds the threshold is a dynamic feature, the quantitative results corresponding to the dynamic feature are statistically analyzed, including: When the property's characteristic of showing records indicates an impact exceeding the threshold, the total number of viewings per day is calculated from the listing date of the sold property to its predicted sale date, which is the quantitative result of the showing records; and / or When the property feature with an impact exceeding the threshold is an interview record, the sum of the number of offline conversations per day from the listing date of the sold property to the predicted transaction date of the sold property is calculated to obtain the total number of offline conversations, which is the quantitative result of the interview record.

8. The information push method according to claim 1, characterized in that, Calculate the feature value of the target property with respect to the first feature, and determine the predicted transaction date of the target property, including: The duration analysis model is invoked to analyze the property characteristics, including the feature values ​​of the first feature, to generate the listing duration of the target property; and The predicted transaction date is determined by overlaying the listing date of the target property with the listing duration.

9. An electronic device, characterized in that, include: The memory stores execution instructions; as well as A processor that executes the execution instructions stored in the memory, causing the processor to perform the information push method according to any one of claims 1 to 8.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the information push method according to any one of claims 1 to 8.

Images