Abnormal housing resource detection method and device, medium and program product

Abstract

The embodiment of the invention relates to an abnormal housing resource detection method and device, a medium and a program product. The method comprises the following steps: acquiring an expected resource transfer amount of a to-be-detected housing resource; on the basis of an anomaly detection strategy corresponding to at least one detection dimension, anomaly detection of the expected resource transfer amount is carried out on the to-be-detected housing resource, and an initial abnormal housing resource on the detection dimension is determined; wherein the detection dimensionality is the dimensionality except the dimensionality related to the resource transfer amount after the house ownership transfer is completed; and performing data cleaning based on each initial abnormal housing resource, and determining a target abnormal housing resource. According to the technical scheme, the accuracy of determining the abnormal housing resources is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to an abnormal housing source detection method, device, medium and program product. Background Art

[0002] Detecting properties with abnormal resource transfer volumes among numerous properties will help effectively protect the rights and interests of consumers.

[0003] The current method of detecting properties with abnormal resource transfer volume is usually to use historical transaction information and compare it with the current expected resource transfer volume of the property to identify abnormal properties. However, in some scenarios, there may not be a large amount of property transaction information. In this way, the reference information of historical transaction information is relatively scarce, resulting in the inability to identify abnormal properties, or the identified abnormal properties are not accurate enough. Summary of the invention

[0004] In order to solve the above technical problems, the present disclosure provides a method, device, medium and program product for detecting abnormal housing sources.

[0005] The present disclosure provides a method for detecting abnormal housing sources, the method comprising:

[0006] Get the expected resource transfer volume of the property to be tested;

[0007] Based on the anomaly detection strategy corresponding to at least one detection dimension, an anomaly detection of the expected resource transfer amount is performed on the detected house source, and the initial abnormal house source on the detection dimension is determined; wherein the detection dimension is a dimension other than the dimension related to the resource transfer amount completed by the transfer of house ownership;

[0008] Perform data cleaning based on each initial abnormal listing and determine the target abnormal listing.

[0009] The present disclosure also provides an abnormal housing source detection device, the device comprising:

[0010] An acquisition module is used to acquire the expected resource transfer amount of the house to be detected;

[0011] A processing module, configured to perform anomaly detection of expected resource transfer amount on the detected house source based on an anomaly detection strategy corresponding to at least one detection dimension, and determine an initial abnormal house source on the detection dimension; wherein the detection dimension is a dimension other than a dimension related to the resource transfer amount completed by the transfer of house ownership;

[0012] The determination module is used to perform data cleaning based on each initial abnormal housing source and determine the target abnormal housing source.

[0013] An embodiment of the present disclosure further provides an electronic device, which includes: a processor and a memory; the processor is used to execute the abnormal housing source detection method described in any embodiment of the present disclosure by calling the program or instruction stored in the memory.

[0014] The embodiments of the present disclosure also provide a computer-readable storage medium, wherein the computer-readable storage medium stores a program or an instruction, wherein the program or the instruction enables a computer to execute the abnormal housing source detection method described in any embodiment of the present disclosure.

[0015] The embodiments of the present disclosure also provide a computer program product, which is used to execute the abnormal housing source detection method described in any embodiment of the present disclosure.

[0016] The abnormal house source detection method, device, medium and program product provided by the embodiments of the present disclosure can perform abnormal detection of the expected resource transfer amount on the house source to be detected through the abnormal detection strategy corresponding to the dimension other than the relevant dimension of the resource transfer amount completed by the transfer of house ownership, determine the initial abnormal house source on the detection dimension, and perform data cleaning based on each initial abnormal house source to determine the target abnormal house source. In other words, the embodiments of the present disclosure can determine the target abnormal house source in the house source to be detected through multiple detection dimensions other than the relevant dimension of the historical transaction price. In this way, the problem of being unable to determine the abnormal house source when the reference information of the historical transaction price is relatively scarce is solved, and the accuracy of determining the abnormal house source is improved. In addition, it is also possible to consider the impact of multiple detection dimensions on the detection results, and fuse the detection results of multiple detection dimensions to obtain the final target abnormal house source, further improving the accuracy of abnormal house source detection. BRIEF DESCRIPTION OF THE DRAWINGS

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

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

[0019] Figure 1 A schematic diagram of an application scenario of an abnormal housing source detection method provided by an embodiment of the present disclosure;

[0020] Figure 2 One of the flowcharts of the abnormal house source detection method provided in the embodiment of the present disclosure;

[0021] Figure 3A second flowchart of the abnormal housing source detection method provided in the embodiment of the present disclosure;

[0022] Figure 4 The third flowchart of the abnormal house source detection method provided in the embodiment of the present disclosure;

[0023] Figure 5 A fourth flowchart of the abnormal housing source detection method provided in an embodiment of the present disclosure;

[0024] Figure 6 A fifth flowchart of the abnormal housing source detection method provided in the embodiment of the present disclosure;

[0025] Figure 7 A visualization view corresponding to a cluster analysis result provided in an embodiment of the present disclosure;

[0026] Figure 8 Flow chart 6 of the abnormal house source detection method provided in the embodiment of the present disclosure;

[0027] Fig. 9 Flow chart 7 of the abnormal house source detection method provided in the embodiment of the present disclosure;

[0028] Fig.10 A schematic diagram of the structure of an abnormal housing source detection device provided in an embodiment of the present disclosure;

[0029] Fig.11 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0030] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described in detail below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0031] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0032] In order to protect the rights and interests of consumers, determining the authenticity of the transfer volume of housing resources is an important task. In the process of identifying real housing sources (that is, housing sources with normal resource transfer volumes), the real and reasonable expected resource transfer volumes can serve as an important reference dimension.

[0033] Currently, historical transaction information is usually used to identify properties with abnormal expected resource transfer volumes. However, in some scenarios, for example, when there is no large amount of historical transaction information as reference information, it is impossible to identify abnormal properties. Secondly, in actual business scenarios, there are various problems of resource transfer volume differences caused by input errors by relevant staff, such as errors in inputting area and total resource transfer volume. There are no targeted detection measures for these anomalies. In addition, some properties may also have no historical expected resource transfer volume information for the same residential area and the same room for reference, which further increases the difficulty of determining abnormal expected resource transfer volumes.

[0034] In view of the above problems, the disclosed embodiment proposes a method for detecting abnormal housing sources, including obtaining the expected resource transfer amount of the housing source to be detected, and based on the abnormal detection strategy corresponding to at least one detection dimension, performing abnormal detection of the expected resource transfer amount on the housing source to be detected, and determining the initial abnormal housing source on the detection dimension. Among them, the detection dimension is a dimension other than the relevant dimension of the resource transfer amount completed by the transfer of housing ownership. Finally, data cleaning is performed based on each initial abnormal housing source to determine the target abnormal housing source. The disclosed embodiment can perform abnormal detection of the expected resource transfer amount on the housing source to be detected based on multiple dimensions, rather than just relying on historical transaction information, or historical expected resource transfer amount to detect abnormal housing sources. In this way, whether it is the lack of historical transaction information, the lack of historical expected resource transfer amount, or the input error of relevant staff, abnormal detection of expected resource transfer amount can be performed on multiple dimensions to determine abnormal housing sources, thereby improving the flexibility and accuracy of determining abnormal housing sources.

[0035] The abnormal housing source detection method provided by the embodiments of the present disclosure can be applied to the situation of abnormal detection of housing sources with abnormal listing prices when there is a lack of historical transaction prices. Figure 1 This is an application scenario diagram of an abnormal housing source detection method provided by an embodiment of the present disclosure. Figure 1 As shown, the application scenario includes a terminal 102 and a server 104. The terminal 102 can communicate with the server 104 through a network. The terminal 102 can include but is not limited to various personal computers, laptops, smart phones, tablet computers, portable wearable devices, etc. The server 104 can be implemented as an independent server or a server cluster composed of multiple servers.

[0036] It should be noted that the abnormal housing source detection method provided in the embodiment of the present application can be executed by an electronic device, and the electronic device can be a server 104, a terminal 102, or a device integrating the functions of a server 104 and a terminal 102, and the present application does not limit this.

[0037] The abnormal housing source detection method provided by the embodiments of the present disclosure can be executed by an abnormal housing source detection device, which can be implemented by software and / or hardware. The device can be integrated into an electronic device with an abnormal housing source detection function, such as a mobile phone, a PDA, a tablet computer, a laptop computer or a desktop computer.

[0038] Figure 2 is a flow chart of an abnormal house source detection method provided by an embodiment of the present disclosure. Figure 1 , the abnormal house source detection method specifically includes:

[0039] S110: Obtain the expected resource transfer amount of the house to be detected.

[0040] Among them, the properties to be tested are properties that need to be tested for abnormal expected resource transfer volume. They can be valid properties that already exist in the current period, or newly listed properties that have been recently added. The expected resource transfer volume of the properties to be tested refers to the number of resources that need to be transferred during the transaction process of the properties to be tested displayed on the trading platform, which can also be understood as the listing price of the properties to be tested.

[0041] Specifically, the abnormal housing source detection device may obtain the expected resource transfer amount of the housing source to be detected by obtaining the expected resource transfer amount of the housing source to be detected input by the user; or it may be to grab the expected resource transfer amount of the housing source corresponding to the housing source identifier from the Internet website according to the housing source identifier input by the user as the expected resource transfer amount of the housing source to be detected, and this application does not limit this. Each housing source has housing source information that characterizes the housing source situation. For housing sources for sale, the housing source information at least includes the expected resource transfer amount. Therefore, the expected resource transfer amount of the housing source to be detected can be extracted from the housing source information.

[0042] In some embodiments, the house to be checked may be a batch of houses or a single house.

[0043] S120. Based on an anomaly detection strategy corresponding to at least one detection dimension, perform an anomaly detection on the expected resource transfer amount for the properties to be detected, and determine the initial abnormal properties in the detection dimension.

[0044] Among them, the detection dimension is a dimension other than the relevant dimension of the amount of resource transfer completed by the transfer of house ownership. The anomaly detection strategy can be an algorithm or method configured in advance for each detection dimension to detect anomalies in the expected amount of resource transfer, such as a statistical algorithm, a clustering algorithm, a machine learning method, etc. One detection dimension can correspond to at least one anomaly detection strategy. The amount of resource transfer completed by the transfer of house ownership refers to the amount of resources transferred when the house transaction is completed, which can also be understood as the transaction price of the house.

[0045] Specifically, the abnormal housing source detection device can call a pre-configured abnormal detection strategy through a preset interface, and perform abnormal detection of the expected resource transfer amount on the housing source to be detected in each detection dimension, so as to determine the initial abnormal housing source corresponding to the housing source to be detected in each dimension.

[0046] In some embodiments, the anomaly detection strategy can be generated by statistical analysis, cluster analysis or data processing of machine learning models on the housing information of historical housing sources. The historical housing sources are a full set of housing sources or a valid set of housing sources; the housing source information includes at least one of the following: the housing source attribute characteristics, the regional characteristics of the area where the housing source is located, and the housing source access behavior characteristics; the housing source attribute characteristics may include the expected unit resource transfer amount of the housing source, the expected resource transfer amount of the housing source, etc.; the regional characteristics of the area where the housing source is located may include the longitude and latitude of the community where the housing source is located, the area of ​​the community, etc.

[0047] Specifically, the method of generating anomaly detection strategies by performing statistical analysis on historical housing information may be to calculate the statistical unit price of historical housing based on the housing information, and generate anomaly detection strategies based on the statistical unit price and a pre-set first difference threshold; the method of generating anomaly detection strategies by performing cluster analysis on historical housing information may be to cluster historical housing based on housing attribute features in the housing information to generate anomaly detection strategies; the method of generating anomaly detection strategies by performing data processing of a machine learning model on historical housing information may be to train a machine learning model using housing information to generate a machine learning model that can predict the unit price of housing, and then continue to generate anomaly detection strategies using the trained machine learning model and a pre-set second difference threshold.

[0048] In some embodiments, the detection dimension may be expanded according to actual conditions. For example, the detection dimension may be multiple dimensions related to the expected resource transfer amount.

[0049] In some embodiments, the detection dimension may include at least one of the statistical dimension of unit resource transfer, the clustering dimension of multi-dimensional house source features, the prediction dimension of resource transfer, and the house source access flow dimension. The unit resource transfer can be understood as the average unit price of the corresponding house in the area where the house to be detected is located. The multi-dimensional house source features include one or more features that can affect the house source price, which can be determined according to the actual situation of the area where the house to be detected is located.

[0050] In some embodiments, the resource transfer amount can be understood as price. The prediction in the prediction dimension of the resource transfer amount is a prediction of the price of each house to be tested, which can be a prediction of the average unit price of the houses to be tested, or a prediction of the total price of the houses to be tested.

[0051] In some embodiments, the listing access flow in the listing access flow dimension can be understood as the number of times the listing is viewed. Since listings with more views are usually priced lower, and listings with fewer views are usually priced higher, determining the initial abnormal listings by the number of times the listings are viewed can more conveniently screen out listings with abnormal prices to be detected.

[0052] S130: Perform data cleaning based on each initial abnormal housing source and determine the target abnormal housing source.

[0053] Specifically, after obtaining the initial abnormal housing sources corresponding to the housing sources to be detected in each dimension, different fusion methods can be used to determine the target abnormal housing sources according to the actual situation corresponding to the business.

[0054] In some embodiments, the data cleaning method may include at least one of deduplication, filtering in combination with business rules, and filtering using statistical methods. Among them, the business rules may be pre-configured algorithms or methods for different businesses, such as threshold determination methods, statistical algorithms, etc. In this way, when there are no other special requirements for the business, direct fusion after deduplication can be used, or the target abnormal housing source can be determined by fusion using statistical methods. When the business requires fusion according to rules, the target abnormal housing source is determined by fusion in combination with business rules.

[0055] The deduplication data cleaning method may be to first remove duplicate listings from the initial abnormal listings corresponding to different dimensions, and then determine the remaining initial abnormal listings as target abnormal listings. For example, when the initial abnormal listings are as shown in Table 1 below, the initial abnormal listing A that is repeated in the first dimension, the second dimension, and the third dimension may be removed to obtain target abnormal listings including A, B, C, and D.

[0056] Table 1

[0057]

[0058] The data cleaning method of filtering by statistical methods is to use the pre-provided statistical methods to count the initial abnormal housing sources and determine the target abnormal housing sources. The statistical methods may include, for example, counting the number of housing source appearances, percentile analysis based on housing source distribution, etc. For example, when the initial abnormal housing sources are as shown in Table 1 above, and the statistical method is to count the initial abnormal housing sources with a number of appearances greater than 1, since B, C, and D only appear once, and A appears 3 times, which is greater than 1, it is determined that the target abnormal housing sources include A.

[0059] The data cleaning method combined with business rule filtering is to filter the initial abnormal housing sources according to the rules provided in the business requirements to obtain the target abnormal housing sources. For example, when the initial abnormal housing sources are as shown in Table 1 above, and the business rule is that the initial abnormal housing sources with a listing price greater than 8 million, the target abnormal housing sources are determined to be packages B and D.

[0060] The above technical scheme of the disclosed embodiment can detect the expected resource transfer amount of the detected house source through the corresponding abnormal detection strategy in dimensions other than the relevant dimension of the resource transfer amount completed by the transfer of house ownership, determine the initial abnormal house source on the detection dimension, and perform data cleaning based on each initial abnormal house source to determine the target abnormal house source. In other words, the disclosed embodiment can determine the target abnormal house source among the detected houses through multiple detection dimensions other than the relevant dimension of the historical transaction price. In this way, the problem of being unable to determine the abnormal house source when the reference information of the historical transaction price is relatively scarce is solved, and the accuracy of determining the abnormal house source is improved. In addition, it is also possible to consider the impact of multiple detection dimensions on the detection results, and fuse the detection results of multiple detection dimensions to obtain the final target abnormal house source, further improving the accuracy of abnormal house source detection.

[0061] Figure 3 is a flow chart of another abnormal house source detection method provided by an embodiment of the present disclosure. In the case where the detection dimension is the statistical dimension of the unit resource transfer amount, step S120 is further optimized. The explanations of the terms that are the same or corresponding to the above embodiments are not repeated here. Figure 3 , the abnormal house source detection method includes:

[0062] S210: Obtain the expected resource transfer amount of the house to be detected.

[0063] S221. Determine the statistical unit resource transfer amount at a preset granularity.

[0064] Among them, the statistical unit resource transfer amount is obtained by statistically processing the unit resource transfer amount of the full set of housing sources according to the preset granularity. The preset granularity can be the granularity of the community corresponding to the housing source to be detected, or the granularity of the housing source in the community corresponding to the housing source to be detected with the same number of rooms as the housing source to be detected, or the granularity of the housing source in the community corresponding to the housing source to be detected with the same number of rooms as the housing source to be detected and the same orientation, etc., that is, the preset granularity can be the community granularity, or the community room granularity, or the community housing source orientation granularity, etc. The full set of housing sources includes valid entrusted housing sources and invalid entrusted housing sources. Valid entrusted housing sources refer to housing sources with true information and can be traded; invalid entrusted housing sources refer to housing sources that cannot be traded because of false information or transactions.

[0065] Specifically, the method for determining the statistical unit resource transfer amount at the preset granularity can be real-time statistics by the abnormal housing source detection device; or statistics can be made in advance, and then when used, the abnormal housing source detection device directly calls the statistical unit resource transfer amount at the preset granularity. The specific statistical method can be selected according to the hardware capability of the abnormal housing source detection device.

[0066] Exemplarily, the statistical unit resource transfer amount at the preset granularity can be represented by the following Table 2, where in cell A, the statistical unit resource transfer amount at the cell granularity is 9972 yuan. At the cell room granularity, when the number of rooms is 1, the statistical unit resource transfer amount is 6000 yuan; when the number of rooms is 2, the statistical unit resource transfer amount is 10622 yuan; when the number of rooms is 3, the statistical unit resource transfer amount is 9437 yuan; when the number of rooms is 4, the statistical unit resource transfer amount is 10769 yuan.

[0067] Table 2

[0068]

[0069] In some embodiments, when the houses to be detected are a batch of houses, the statistical unit resource transfer amount at a preset granularity is determined for each house to be detected.

[0070] S222. Based on the expected resource transfer amount of the house to be detected, determine the expected unit resource transfer amount of the house to be detected at a preset granularity, and based on the expected unit resource transfer amount at the preset granularity and the statistical unit resource transfer amount, determine the first resource transfer amount difference of the house to be detected.

[0071] First, based on the expected resource transfer amount of the house to be detected, the expected unit resource transfer amount of the house to be detected at a preset granularity is determined. The expected resource transfer amount of the house to be detected can be understood as the listing unit price of the house to be detected.

[0072] Specifically, before determining the expected unit resource transfer amount of the house to be detected at a preset granularity based on the expected resource transfer amount of the house to be detected, it is necessary to first determine the area of ​​the house to be detected, and then calculate the expected unit resource transfer amount of the house to be detected at the preset granularity according to the formula A÷B=C. Among them, C is used to represent the expected unit resource transfer amount of the house to be detected at the preset granularity, A is used to represent the expected resource transfer amount of the house to be detected, and B is used to represent the area of ​​the house to be detected.

[0073] Secondly, based on the expected unit resource transfer amount and the statistical unit resource transfer amount at a preset granularity, the first resource transfer amount difference of the house to be detected is determined.

[0074] Specifically, the first resource transfer amount difference of the house to be detected can be determined by directly making a difference or calculating a deviation.

[0075] When the first resource transfer amount difference is determined by direct difference, the difference between the expected unit resource transfer amount and the statistical unit resource transfer amount at the preset granularity, or the difference between the statistical unit resource transfer amount and the expected unit resource transfer amount at the preset granularity, can be directly determined as the first resource transfer amount difference of the house to be detected. It should be noted that in the detection process of the same abnormal house, when calculating the first resource transfer amount difference, the objects corresponding to the subtrahend and the minuend should remain consistent.

[0076] When determining the first resource transfer amount difference of the house to be detected by calculating the deviation, the first resource transfer amount difference may be calculated according to the formula (DE)÷E=F, where D is used to represent the expected unit resource transfer amount, E is used to represent the statistical unit resource transfer amount, and F is used to represent the first resource transfer amount difference.

[0077] In some embodiments, when the houses to be detected are a batch of houses, the first resource transfer amount difference of each house to be detected is determined respectively.

[0078] S223. If the first resource transfer amount difference is greater than the first difference threshold, the house to be detected is determined as the initial abnormal house in the statistical dimension of the unit resource transfer amount.

[0079] The first difference threshold is preset, for example, the first difference threshold may be a default value or a preset value.

[0080] Specifically, when the first resource transfer amount difference is greater than the first difference threshold, the expected unit resource transfer amount of the house to be detected is determined to be beyond a pre-set normal range, and the house to be detected will be determined as the initial abnormal house in the statistical dimension of the unit resource transfer amount.

[0081] In some embodiments, when the houses to be detected are a batch of houses, it is determined whether each house to be detected is an initial abnormal house in the statistical dimension of unit resource transfer amount.

[0082] In some embodiments, if the first resource transfer amount difference is less than or equal to the first difference threshold, the abnormal housing source detection device does not perform any processing.

[0083] S230: Perform data cleaning based on each initial abnormal housing source and determine the target abnormal housing source.

[0084] Specifically, the abnormal housing source detection device can use at least one of the above-mentioned deduplication, filtering in combination with business rules, and filtering using statistical methods to clean the data of each initial abnormal housing source to obtain the target abnormal housing source.

[0085] The above technical solution of the disclosed embodiment can determine whether the house to be detected is an initial abnormal house by comparing the statistical unit resource transfer volume of the entire house set with the expected unit resource transfer volume of the house to be detected. This solves the problem that the abnormal house cannot be determined due to the lack of a large amount of historical transaction information as reference information, avoids the reliance on historical transaction prices when determining the initial abnormal house, and improves the flexibility of determining abnormal houses.

[0086] In some embodiments, Figure 4 As shown, the method of determining the statistical unit resource transfer amount at a preset granularity includes the following steps:

[0087] S2211. Filter the entire set of housing sources to generate a filtered set of housing sources.

[0088] Specifically, the full set of housing sources can be filtered according to the preset filtering rules to generate a filtered housing source set. The filtering rules at least include the filtering rules for the housing sources corresponding to the preset granularity, so as to filter out the filtered housing source set at the corresponding granularity. For example, when the preset granularity is the cell granularity, the filtering rules should include the filtering rules corresponding to the cell; when the preset granularity is the cell room granularity, the filtering rules should include the filtering rules corresponding to the cell and the number of rooms. For another example, the filtering rule corresponding to the number of rooms can be that the number of rooms is between 1 and 6.

[0089] In some embodiments, the filtering rules may also include screening rules corresponding to the listing price, housing area, etc., to filter out the listing prices and housing areas of the full housing collection, and form a filtered housing collection to avoid the influence of abnormal housing on subsequent statistical results. For example, the screening rule for the listing price may be between 1,000 and 100,000 yuan / square meter; the screening rule for the housing area may be between 10 and 300 square meters.

[0090] S2212: Determine the weight value of the house to be counted based on the property attribute information of any house to be counted in the filtered house set, and determine the expected unit resource transfer amount of the house to be counted at a preset granularity based on the expected resource transfer amount of the house to be counted.

[0091] First, based on the property attribute information of any property to be counted in the filtered property set, the weight value of the property to be counted is determined. The property attribute information of the property includes at least one of the entrustment status, display status, modification times and modification time of the property to be counted. The entrustment status includes a valid entrustment status and an invalid entrustment status, the display status includes an external display status and a non-external display status, the modification times of the resource transfer amount can be understood as the price modification times of the property to be counted, and the modification time can be understood as the price time of the property to be counted.

[0092] Specifically, the entrustment weight value, display weight value, modification number weight value and modification time weight value of the house to be counted are determined based on the entrustment status, display status, modification times and modification time of the resource transfer amount of the house to be counted.

[0093] Based on the entrustment status of the houses to be counted, the method for determining the entrustment weight value of the houses to be counted can be to determine the entrustment weight value of the houses to be counted whose entrustment status is a valid entrustment status as a first value (such as 2); and to determine the entrustment weight value of the houses to be counted whose entrustment status is an invalid entrustment status as a second value (such as 1).

[0094] Based on the display status of the houses to be counted, the display weight values ​​of the houses to be counted can be determined as follows: the display weight values ​​of the houses to be counted whose display status is the external display status can be determined as a third value (such as 1.5); and the display weight values ​​of the houses to be counted whose display status is the non-external display status can be determined as a fourth value (such as 1).

[0095] Based on the number of modification times of the resource transfer amount of the to-be-counted house, the method of determining the modification times weight value of the to-be-counted house can be to determine the modification times weight value of the to-be-counted house whose number of modification times of the resource transfer amount is greater than or equal to the number threshold as the fifth value (such as 1.5), and to determine the modification times weight value of the to-be-counted house whose number of modification times of the resource transfer amount is less than the number threshold as the sixth value (such as 1). The number threshold is preset, such as a default value, or a value set by relevant personnel according to actual conditions. For another example, the number threshold is 2.

[0096] Based on the modification time of the listings to be counted, the modification time weight value of the listings to be counted can be determined by first calculating the modification time difference between the last modification time and the current time, and then determining the modification time weight value according to the modification time difference and the corresponding time threshold. For example, the modification time weight value of the listings to be counted whose modification time difference is less than the time threshold can be determined as the seventh value; the modification time weight value of the listings to be counted whose modification time difference is greater than or equal to the time threshold can be determined as the eighth value.

[0097] Secondly, based on the expected resource transfer volume of the houses to be counted, the expected unit resource transfer volume of the houses to be counted at a preset granularity is determined.

[0098] Specifically, the method for determining the expected unit resource transfer amount of the houses to be counted is the same as the method for determining the expected unit resource transfer amount of the houses to be detected in step S222, and will not be repeated here.

[0099] S2213: Perform weighted average processing based on the weight value of each house source to be counted and the expected unit resource transfer amount at a preset granularity to determine the statistical unit resource transfer amount at the preset granularity.

[0100] Specifically, according to the formula Determine the statistical unit resource transfer amount at the preset granularity. Among them, M is used to represent the statistical unit resource transfer amount, si is used to represent the expected unit resource transfer amount of the i-th house to be counted at the preset granularity, ai is used to represent the entrustment weight value of the i-th house to be counted, bi is used to represent the display weight value of the i-th house to be counted, ci is used to represent the modification number weight value of the i-th house to be counted, and di is used to represent the modification time weight value of the i-th house to be counted.

[0101] The above technical solution of the embodiment of the present disclosure performs weighted average processing on the expected resource transfer amount of the housing source to be counted through the property information of the housing source to be counted, and obtains the statistical unit resource transfer amount of the housing source to be counted. In this way, on the one hand, the expected resource transfer amount of the housing source to be counted can be used to calculate the statistical unit resource transfer amount, which provides reference information for the subsequent determination of abnormal housing sources and avoids dependence on historical transaction prices when determining abnormal housing sources. On the other hand, the property information of the housing source can also be added to the calculation of the statistical unit resource transfer amount, making the calculation result more accurate.

[0102] In some embodiments, Figure 5 As shown, the method of determining the modification time weight value of the house to be counted based on the modification time of the resource transfer amount of the house to be counted includes the following steps:

[0103] S1. Determine the modification time difference of the resource transfer amount based on the current time and the last modification time of the resource transfer amount of the house to be counted.

[0104] Specifically, the time length between the current time and the last modification time of the resource transfer amount of the house to be counted is determined as the modification time difference of the resource transfer amount.

[0105] S2. Determine whether the modification time difference is less than or equal to the first time threshold; if the modification time difference is less than or equal to the first time threshold, execute step S3; if the modification time is greater than the first time threshold, execute step S4.

[0106] The first time threshold is preset, for example, a default value, or a value set by relevant personnel according to actual conditions. For another example, the first time threshold is 30 days.

[0107] S3. Determine the preset weight value as the modification time weight value of the house to be counted.

[0108] The preset weight value is a preset value, for example, the preset weight value is 1.

[0109] S4. Determine whether the modification time difference is greater than the first time threshold and less than or equal to the second time threshold; if the modification time difference is greater than the first time threshold and less than or equal to the second time threshold, execute step S5; if the modification time difference is greater than the second time threshold, execute step S6.

[0110] The second time threshold is preset, for example, a default value, or a value set by relevant personnel according to actual conditions. For another example, the second time threshold is 365 days.

[0111] S5. Perform nonlinear processing based on the modification time difference and the first coefficient to determine the modification time weight value of the listing to be counted.

[0112] The first coefficient is preset, for example, a default value, or a value set by relevant personnel according to actual conditions. For another example, the first coefficient is -0.004.

[0113] Specifically, according to the formula Nonlinear processing is performed based on the modification time difference and the first coefficient to determine the modification time weight value of the house to be counted, where d is used to represent the modification time weight value of the house to be counted, l1 is used to represent the first coefficient, and t is used to represent the modification time difference.

[0114] S6. Perform nonlinear processing based on the modification time difference, the first coefficient and the second coefficient to determine the modification time weight value of the house to be counted.

[0115] The second coefficient is preset, for example, a default value, or a value set by relevant personnel according to actual conditions. For another example, the second coefficient is 0.5.

[0116] Specifically, according to the formula Nonlinear processing is performed based on the modification time difference, the first coefficient and the second coefficient to determine the modification time weight value of the house to be counted, where d is used to represent the modification time weight value of the house to be counted, l1 is used to represent the first coefficient, l2 is used to represent the second coefficient, and t is used to represent the modification time difference.

[0117] The above-mentioned technical scheme of the embodiment of the present disclosure determines different modification time weight values ​​for the houses to be counted through the modification time difference of the resource transfer amount of the houses to be counted. Since the modification time difference can reflect the demand of the houses to be counted in the market, different modification time weight values ​​are determined according to different modification time differences. This can better avoid the influence of extreme values ​​on the subsequent calculation of the statistical unit resource transfer amount, thereby improving the accuracy of determining the statistical unit resource transfer amount.

[0118] Figure 6 1 is a flowchart of another abnormal house source detection method provided by an embodiment of the present disclosure. In the case where the detection dimension is a clustering dimension of multi-dimensional house source features, step S120 is further optimized. The explanations of the terms that are the same or corresponding to the above embodiments are not repeated here. Figure 6 , the abnormal house source detection method includes:

[0119] S310: Obtain the expected resource transfer amount of the house to be detected.

[0120] S321. Perform cluster analysis on the first valid housing source set based on a preset clustering model to determine cluster anomalies in the valid housing source set.

[0121] The first valid housing source set is a set of valid entrusted housing sources, and the first valid housing source set includes the housing sources to be detected. In some embodiments, a housing source to be detected and the valid housing sources in the community where it is located are combined into a first valid housing source set. The valid entrusted housing sources refer to the housing sources with real information and can be traded.

[0122] The preset clustering model is a pre-set unsupervised clustering model, which has at least the following characteristics: it can cluster dense data sets of any distribution shape; clustering anomalies can be found while clustering (relatively insensitive to anomalies); the clustering results are not biased, and the algorithm in the model is less affected by preset values ​​such as initial values; there is no need to specify the number of clusters, etc. For example, the preset clustering model can be a DBScan model.

[0123] The multi-dimensional house source features of the preset clustering model include at least one of the house source attribute features, the regional features of the area where the house source is located, and the statistical unit resource transfer amount at the regional granularity. Among them, the house source attribute features at least include the expected unit resource transfer amount of the house source, and may also include the expected resource transfer amount of the house source, the area of ​​the house source, the number of rooms in the house source, the longitude and latitude of the house source, the orientation of the house source, the purpose of the house source (such as commercial housing, residential, etc.), and the difference between the expected unit resource transfer amount of other houses in the same community, etc. The regional features of the area where the house source is located may include the longitude and latitude of the community where the house source is located, the area of ​​the community, etc.

[0124] Specifically, a cluster analysis is performed on the first valid housing source set based on a preset clustering model to determine cluster anomalies in the valid housing source set. Multidimensional housing source features may be input into the preset clustering model, and the preset clustering model may perform cluster analysis on the housing sources in the first valid housing source set according to the multidimensional housing source features to obtain a cluster analysis result.

[0125] For example, Figure 7 A visualization view corresponding to a cluster analysis result provided in an embodiment of the present disclosure is a visualization view including two-dimensional features obtained after dimensionality reduction of the cluster analysis result corresponding to the multi-dimensional property source features, such as Figure 7 As shown in the figure, each cluster point represents a property, the horizontal axis represents one dimension after dimensionality reduction, such as the first dimension, and the vertical axis represents another dimension after dimensionality reduction, such as the second dimension. The visualization view includes cluster normal points (more concentrated) and cluster abnormal points (more dispersed) obtained by clustering analysis based on multi-dimensional property characteristics, so that cluster abnormal points in the valid property set can be determined from the cluster analysis results. In this way, the cluster analysis results can be visualized in two dimensions, making the cluster analysis results more intuitive and easy to understand.

[0126] S322: If the house to be detected is a cluster outlier, the house to be detected is determined as an initial outlier house on the cluster dimension of the multi-dimensional house feature.

[0127] In some embodiments, if the house to be detected does not belong to the cluster outlier point, the abnormal house detection device does not process it.

[0128] S330: Perform data cleaning based on each initial abnormal housing source and determine the target abnormal housing source.

[0129] Specifically, the abnormal housing source detection device can use at least one of the above-mentioned deduplication, filtering in combination with business rules, and filtering using statistical methods to clean the data of each initial abnormal housing source to obtain the target abnormal housing source.

[0130] The above technical solution of the disclosed embodiment can perform cluster analysis on the multi-dimensional house source features of the first valid house source set including the house source to be detected through a preset clustering model, and determine whether the house source to be detected is an initial abnormal house source. This solves the problem of no targeted detection measures for resource transfer volume differences caused by input errors by relevant staff, and improves the flexibility of determining abnormal houses.

[0131] Figure 8 is a flowchart of another abnormal house source detection method provided by an embodiment of the present disclosure. In the case where the detection dimension is the prediction dimension of the resource transfer amount, step S120 is further optimized. The explanations of the terms that are the same or corresponding to the above embodiments are not repeated here. Figure 8, the abnormal house source detection method includes:

[0132] S410: Obtain the expected resource transfer amount of the house to be detected.

[0133] S421. Input the house source information of the house source to be detected into a preset regression model to determine the predicted unit resource transfer amount of the house source to be detected.

[0134] The house source information includes at least one of the house source attribute characteristics, the regional characteristics of the area where the house source is located, and the house source access behavior characteristics. The preset regression model is preset, for example, it can be a CatBoost model, or the preset regression model is obtained by performing regression training in advance using the second valid house source set. The second valid house source set is a set of valid entrusted houses, and the second valid house source set does not include the house to be detected.

[0135] S422. Based on the expected resource transfer amount of the house to be detected, determine the expected unit resource transfer amount of the house to be detected at the house source granularity, and based on the expected unit resource transfer amount and the predicted unit resource transfer amount at the house source granularity, determine the second resource transfer amount difference of the house to be detected.

[0136] Specifically, the method of determining the expected unit resource transfer amount of the house to be detected at the house source granularity based on the expected resource transfer amount of the house to be detected is the same as the method of determining the expected unit resource transfer amount of the house to be detected at the preset granularity based on the expected resource transfer amount of the house to be detected in step S222. The method of determining the difference of the second resource transfer amount of the house to be detected based on the expected unit resource transfer amount and the predicted unit resource transfer amount at the house source granularity is the same as the method of determining the difference of the first resource transfer amount of the house to be detected based on the expected unit resource transfer amount and the statistical unit resource transfer amount at the preset granularity in step S222, and will not be repeated here.

[0137] In some embodiments, when the houses to be detected are a batch of houses, the second resource transfer amount difference of each house to be detected is determined respectively.

[0138] S423: If the second resource transfer amount difference is greater than the second difference threshold, the house to be detected is determined as the initial abnormal house in the prediction dimension of the resource transfer amount.

[0139] The second difference threshold is preset, for example, the second difference threshold may be a default value or a preset value.

[0140] In some embodiments, when the houses to be detected are a batch of houses, the initial abnormal houses of each house to be detected in the prediction dimension of the resource transfer amount are determined respectively.

[0141] In some embodiments, if the second resource transfer amount difference is less than or equal to the second difference threshold, the abnormal housing source detection device does not perform any processing.

[0142] S430: Perform data cleaning based on each initial abnormal housing source and determine the target abnormal housing source.

[0143] Specifically, the abnormal housing source detection device can use at least one of the above-mentioned deduplication, filtering in combination with business rules, and filtering using statistical methods to clean the data of each initial abnormal housing source to obtain the target abnormal housing source.

[0144] The above technical solution of the disclosed embodiment can determine whether the house to be detected is an initial abnormal house by comparing the predicted unit resource transfer of the house to be detected and the expected unit resource transfer amount. This solves the problem that some houses may have no historical expected resource transfer amount information for the same residential area and the same room for reference, which makes it difficult to determine the abnormal expected resource transfer amount, and improves the flexibility of determining abnormal houses.

[0145] Fig. 9 is a flow chart of another abnormal house source detection method provided by an embodiment of the present disclosure. Step S120 and step S130 are further optimized. The explanations of the terms that are the same or corresponding to the above embodiments are not repeated here. Fig. 9 , the abnormal house source detection method includes:

[0146] S510: Obtain the expected resource transfer amount of the house to be detected.

[0147] S520, determining the initial abnormal housing sources in the statistical dimension of unit resource transfer amount, the initial abnormal housing sources in the clustering dimension of multi-dimensional housing source characteristics, the initial abnormal housing sources in the prediction dimension of resource transfer amount, and the initial abnormal housing sources in the housing source access flow dimension.

[0148] Specifically, the method for determining the initial abnormal listings in the statistical dimension of unit resource transfer volume is detailed in steps S221-S223; the method for determining the initial abnormal listings in the clustering dimension of multi-dimensional listing features is detailed in steps S321-S322; the method for determining the initial abnormal listings in the prediction dimension of resource transfer volume is detailed in steps S421-S423; the method for determining the initial abnormal listings in the listing access traffic dimension is detailed in step S120.

[0149] S530, perform data cleaning on the initial abnormal listings on at least two detection dimensions, including the initial abnormal listings on the statistical dimension of unit resource transfer volume, the initial abnormal listings on the clustering dimension of multi-dimensional listing characteristics, the initial abnormal listings on the prediction dimension of resource transfer volume, and the initial abnormal listings on the listing access traffic dimension, to determine the target abnormal listings.

[0150] First, the abnormal housing source detection device can perform data cleaning on any two, any three, or the union of the initial abnormal housing sources on the statistical dimension of unit resource transfer volume, the initial abnormal housing sources on the clustering dimension of multi-dimensional housing source characteristics, the initial abnormal housing sources on the prediction dimension of resource transfer volume, and the initial abnormal housing sources on the housing source access traffic dimension to determine the target abnormal housing sources.

[0151] Secondly, the abnormal housing source detection device can use at least one of the deduplication described in the above step S130, filtering in combination with business rules, and filtering using statistical methods to clean the data of each initial abnormal housing source to obtain the target abnormal housing source.

[0152] The above technical solution of the embodiment of the present disclosure can perform data cleaning on the union of at least two detection dimensions of initial abnormal housing sources in the statistical dimension of unit resource transfer volume, initial abnormal housing sources in the clustering dimension of multi-dimensional housing source features, initial abnormal housing sources in the prediction dimension of resource transfer volume, and initial abnormal housing sources in the housing source access flow dimension to determine the target abnormal housing source. In this way, the influence of multiple detection dimensions on the detection results can be considered, and the detection results of multiple detection dimensions can be integrated to obtain the final target abnormal housing source, further improving the accuracy of abnormal housing source detection.

[0153] Fig.10 The following is a schematic diagram of the structure of an abnormal housing source detection device provided by an embodiment of the present disclosure. Fig.10 As shown, the abnormal house source detection device includes:

[0154] An acquisition module 91 is used to acquire the expected resource transfer amount of the house to be detected;

[0155] A processing module 92 is used to perform anomaly detection of the expected resource transfer amount on the detected house source based on an anomaly detection strategy corresponding to at least one detection dimension, and determine an initial abnormal house source on the detection dimension; wherein the detection dimension is a dimension other than a dimension related to the resource transfer amount completed by the house ownership transfer;

[0156] The determination module 93 is used to perform data cleaning based on each initial abnormal housing source and determine the target abnormal housing source.

[0157] An abnormal house source detection device provided by the embodiment of the present disclosure can detect the abnormal amount of expected resource transfer for the house source to be detected through the dimensions other than the relevant dimensions of the resource transfer amount completed by the transfer of house ownership and the corresponding abnormal detection strategy, determine the initial abnormal house source on the detection dimension, and perform data cleaning based on each initial abnormal house source to determine the target abnormal house source. In other words, the embodiment of the present disclosure can determine the target abnormal house source among the houses to be detected through multiple detection dimensions other than the relevant dimensions of the historical transaction price. In this way, the problem of being unable to determine the abnormal house source when the reference information of the historical transaction price is relatively scarce is solved, and the accuracy of determining the abnormal house source is improved. In addition, the influence of multiple detection dimensions on the detection results can be considered, and the detection results of multiple detection dimensions can be integrated to obtain the final target abnormal house source, further improving the accuracy of abnormal house source detection.

[0158] In some embodiments, the anomaly detection strategy is generated by performing statistical analysis, cluster analysis, or data processing of a machine learning model on historical listings.

[0159] In some embodiments, the detection dimension includes at least one of a statistical dimension of unit resource transfer volume, a clustering dimension of multi-dimensional property characteristics, a prediction dimension of resource transfer volume, and a property access traffic dimension.

[0160] In some embodiments, the processing module 92 is specifically configured to:

[0161] When the detection dimension is a statistical dimension of the unit resource transfer amount, the statistical unit resource transfer amount at a preset granularity is determined; wherein the statistical unit resource transfer amount is obtained by statistically processing the unit resource transfer amount of the full set of housing sources at the preset granularity; the full set of housing sources includes valid entrusted housing sources and invalid entrusted housing sources;

[0162] Based on the expected resource transfer amount of the house to be detected, determine the expected unit resource transfer amount of the house to be detected at a preset granularity, and determine a first resource transfer amount difference value of the house to be detected based on the expected unit resource transfer amount at the preset granularity and the statistical unit resource transfer amount;

[0163] If the first resource transfer amount difference is greater than the first difference threshold, the house to be detected is determined as the initial abnormal house in the statistical dimension of the unit resource transfer amount.

[0164] In some embodiments, the processing module 92 is specifically configured to:

[0165] Filter the entire set of properties to generate a filtered set of properties;

[0166] Based on the property attribute information of any to-be-counted property in the filtered property set, a weight value of the to-be-counted property is determined, and based on the expected resource transfer amount of the to-be-counted property, an expected unit resource transfer amount of the to-be-counted property at a preset granularity is determined;

[0167] A weighted average processing is performed based on the weight value of each house to be counted and the expected unit resource transfer amount at a preset granularity to determine the statistical unit resource transfer amount at the preset granularity.

[0168] In some embodiments, the processing module 92 is specifically used to determine the entrustment weight value, display weight value, modification number weight value and modification time weight value of the house to be counted based on the entrustment status, display status, modification number and modification time of the resource transfer amount of the house to be counted.

[0169] In some embodiments, the processing module 92 is specifically configured to:

[0170] Based on the current time and the last modification time of the resource transfer amount of the listing to be counted, determine the modification time difference of the resource transfer amount;

[0171] If the modification time difference is less than or equal to the first time threshold, the preset weight value is determined as the modification time weight value of the house to be counted;

[0172] If the modification time difference is greater than the first time threshold and less than or equal to the second time threshold, nonlinear processing is performed based on the modification time difference and the first coefficient to determine the modification time weight value of the house to be counted;

[0173] If the modification time difference is greater than the second time threshold, nonlinear processing is performed based on the modification time difference, the first coefficient and the second coefficient to determine the modification time weight value of the house to be counted.

[0174] In some embodiments, the processing module 92 is specifically configured to:

[0175] In the case where the detection dimension is a clustering dimension of a multidimensional property feature, cluster analysis is performed on the first valid property set based on a preset clustering model to determine cluster anomalies in the valid property set; wherein the first valid property set is a set of valid entrusted properties, and the first valid property set includes properties to be detected; the multidimensional property feature of the preset clustering model includes at least one of property attribute features, regional features of the area where the property is located, and a statistical unit resource transfer amount at a regional granularity;

[0176] If the house to be detected is a cluster outlier, the house to be detected is determined as the initial outlier house on the cluster dimension of the multi-dimensional house feature.

[0177] In some embodiments, the processing module 92 is specifically configured to:

[0178] In the case where the detection dimension is a prediction dimension of resource transfer amount, inputting the house source information of the house source to be detected into a preset regression model to determine the predicted unit resource transfer amount of the house source to be detected; wherein the house source information includes at least one of a house source attribute feature, a regional feature of a region where the house source is located, and a house source access behavior feature;

[0179] Based on the expected resource transfer amount of the house to be detected, determine the expected unit resource transfer amount of the house to be detected at the house source granularity, and based on the expected unit resource transfer amount and the predicted unit resource transfer amount at the house source granularity, determine the second resource transfer amount difference of the house to be detected;

[0180] If the second resource transfer amount difference is greater than the second difference threshold, the house to be detected is determined as the initial abnormal house in the prediction dimension of the resource transfer amount.

[0181] In some embodiments, the preset regression model is obtained by performing regression training in advance using a second valid housing source set; the second valid housing source set is a set consisting of valid entrusted housing sources, and the second valid housing source set does not include housing sources to be detected.

[0182] The abnormal housing source detection device provided in the embodiments of the present disclosure can execute the abnormal housing source detection method provided in any embodiment of the present disclosure, and has the corresponding functional modules and beneficial effects of the execution method.

[0183] It is worth noting that in the embodiment of the above-mentioned abnormal housing source detection device, the various modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the various functional modules are only for the convenience of distinguishing each other, and are not used to limit the protection scope of the present disclosure.

[0184] Fig.11 The structure diagram of an electronic device provided by the embodiment of the present disclosure is shown in FIG. Fig.11 As shown, the electronic device 400 includes a processor 401, a memory 402, an input device 403 and an output device 404, and these components are interconnected through a bus system and / or other forms of connection mechanisms (not shown). The number of the processor 401 and the memory 402 can be one or more, Figure 4 In the figure, a processor 401 and a memory 402 are taken as an example.

[0185] The processor 401 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 400 to perform desired functions.

[0186] The memory 402 may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. In some embodiments, the memory 402 may further include a memory remotely arranged relative to the processor 401, and these remote memories may be connected to the electronic device 400 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. One or more computer programs or instructions may be stored in the above-mentioned memory 402, and the processor 401 may run these programs or instructions to implement the abnormal housing source detection method and / or other desired functions of any embodiment of the present disclosure. Various contents such as housing source information, multi-dimensional housing source features, etc. may also be stored in the above-mentioned memory 402.

[0187] The input device 403 may include, for example, a keyboard, a mouse, etc. The output device 404 may output various information to the outside, including the determined target abnormal housing sources, a visual view of the cluster analysis results, etc. The output device 404 may include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.

[0188] Understandably, for simplicity, Fig.11 Only some of the components related to the present disclosure in the electronic device 400 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device 400 may also include any other appropriate components.

[0189] In addition to the above-mentioned methods and devices, the abnormal housing source detection method in any embodiment of the present disclosure can also be implemented as a computer software program. For example, the present disclosure also includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a memory. When the computer program is executed by a processor, the processor executes the abnormal housing source detection method provided in any embodiment of the present disclosure.

[0190] The computer program product may be written in any combination of one or more programming languages ​​to write program code for performing the operations of the disclosed embodiments, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0191] In addition, an embodiment of the present disclosure further provides a computer-readable storage medium having a computer program or instructions stored thereon. When the computer program or instructions are executed by a processor, the processor executes the abnormal housing source detection method provided by an embodiment of the present disclosure.

[0192] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0193] It should be noted that the terms used in the present disclosure are only for describing specific embodiments, rather than limiting the scope of the present application. As shown in the present disclosure specification and claims, unless the context clearly indicates an exception, the words "one", "a", "a kind of" and / or "the" do not specifically refer to the singular, but may also include the plural. The term "and / or" includes any and all combinations of one or more related listed items. Relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. The term "include", "comprise" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method or device. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method or device including the elements.

[0194] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting abnormal housing sources, characterized in that: include: Get the expected resource transfer volume of the property to be tested; Based on an anomaly detection strategy corresponding to at least one detection dimension, an anomaly detection of the expected resource transfer amount is performed on the house to be detected, and an initial abnormal house on the detection dimension is determined; wherein the detection dimension is a dimension other than a dimension related to the resource transfer amount completed by the transfer of house ownership; Data cleaning is performed based on each of the initial abnormal listings to determine a target abnormal listing.

2. The method according to claim 1, characterized in that The anomaly detection strategy is generated by performing statistical analysis, cluster analysis or data processing of a machine learning model on the listing information of historical listings.

3. The method according to claim 2, characterized in that The detection dimension includes at least one of a statistical dimension of unit resource transfer volume, a clustering dimension of multi-dimensional property characteristics, a prediction dimension of resource transfer volume, and a property access traffic dimension.

4. The method according to claim 3, characterized in that In the case where the detection dimension is the statistical dimension of the unit resource transfer amount, based on the anomaly detection strategy corresponding to the statistical dimension of the unit resource transfer amount, anomaly detection of the expected resource transfer amount is performed on the house to be detected, and an initial abnormal house on the detection dimension is determined, including: Determine a statistical unit resource transfer amount at a preset granularity; wherein the statistical unit resource transfer amount is obtained by statistically processing the unit resource transfer amount of the full set of housing sources according to the preset granularity; the full set of housing sources includes valid entrusted housing sources and invalid entrusted housing sources; Based on the expected resource transfer amount of the house to be detected, determining the expected unit resource transfer amount of the house to be detected at the preset granularity, and based on the expected unit resource transfer amount at the preset granularity and the statistical unit resource transfer amount, determining a first resource transfer amount difference value of the house to be detected; If the first resource transfer amount difference is greater than a first difference threshold, the house to be detected is determined as the initial abnormal house in the statistical dimension of the unit resource transfer amount.

5. The method according to claim 4, characterized in that The determining of the statistical unit resource transfer amount at a preset granularity includes: Filtering the full set of listings to generate a filtered set of listings; Based on the property attribute information of any to-be-counted property in the filtered property set, a weight value of the to-be-counted property is determined, and based on the expected resource transfer amount of the to-be-counted property, an expected unit resource transfer amount of the to-be-counted property at the preset granularity is determined; A weighted average processing is performed based on the weight value of each of the houses to be counted and the expected unit resource transfer amount at the preset granularity to determine the statistical unit resource transfer amount at the preset granularity.

6. The method according to claim 5, characterized in that The determining the weight value of any house to be counted based on the house attribute information of any house to be counted in the filtered house set includes: The entrustment weight value, display weight value, modification number weight value and modification time weight value of the house to be counted are determined based on the entrustment status, display status, modification times and modification time of the resource transfer amount of the house to be counted.

7. The method according to claim 6, characterized in that Determining a modification time weight value of the house to be counted based on the modification time of the resource transfer amount of the house to be counted includes: Determine the modification time difference of the resource transfer amount based on the current time and the last modification time of the resource transfer amount of the house to be counted; If the modification time difference is less than or equal to the first time threshold, the preset weight value is determined as the modification time weight value of the house to be counted; If the modification time difference is greater than the first time threshold and less than or equal to the second time threshold, nonlinear processing is performed based on the modification time difference and the first coefficient to determine the modification time weight value of the house to be counted; If the modification time difference is greater than the second time threshold, nonlinear processing is performed based on the modification time difference, the first coefficient and the second coefficient to determine the modification time weight value of the house to be counted.

8. The method according to claim 3, characterized in that In the case where the detection dimension is a clustering dimension of the multi-dimensional property feature, based on an anomaly detection strategy corresponding to the clustering dimension of the multi-dimensional property feature, an anomaly detection of the expected resource transfer amount is performed on the property to be detected, and an initial abnormal property on the detection dimension is determined, including: Performing cluster analysis on a first valid housing source set based on a preset clustering model to determine cluster outliers in the valid housing source set; wherein the first valid housing source set is a set of valid entrusted housing sources, and the first valid housing source set includes the housing source to be detected; the multi-dimensional housing source features of the preset clustering model include at least one of housing source attribute features, regional features of the area where the housing source is located, and statistical unit resource transfer volume at regional granularity; If the house to be detected belongs to the cluster outlier, the house to be detected is determined as the initial outlier house on the cluster dimension of the multi-dimensional house feature.

9. The method according to claim 3, characterized in that: In the case where the detection dimension is the prediction dimension of the resource transfer amount, based on the anomaly detection strategy corresponding to the prediction dimension of the resource transfer amount, anomaly detection of the expected resource transfer amount is performed on the house to be detected, and an initial abnormal house on the detection dimension is determined, including: Inputting the house source information of the house source to be detected into a preset regression model to determine the predicted unit resource transfer amount of the house source to be detected; wherein the house source information includes at least one of the house source attribute characteristics, the regional characteristics of the area where the house source is located, and the house source access behavior characteristics; Based on the expected resource transfer amount of the house to be detected, determining the expected unit resource transfer amount of the house to be detected at the house source granularity, and based on the expected unit resource transfer amount at the house source granularity and the predicted unit resource transfer amount, determining a second resource transfer amount difference value of the house to be detected; If the second resource transfer amount difference is greater than a second difference threshold, the house to be detected is determined as the initial abnormal house in the predicted dimension of the resource transfer amount.

10. The method according to claim 9, characterized in that The preset regression model is obtained by performing regression training in advance using a second valid housing source set; the second valid housing source set is a set of valid entrusted housing sources, and the second valid housing source set does not include the housing source to be detected.

11. An electronic device, characterized in that: The electronic device comprises: Processor and memory; The processor is used to execute the abnormal housing source detection method as described in any one of claims 1 to 10 by calling the program or instruction stored in the memory.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program or instruction, and the program or instruction enables a computer to execute the abnormal housing source detection method as described in any one of claims 1 to 10.

13. A computer program product, characterized in that The computer program product is used to implement the abnormal housing source detection method as described in any one of claims 1 to 10.

Images