House resource evaluation model optimization method and device, medium and computer program product
By obtaining and optimizing the listing feature information of the property valuation model, the problem of lag in the property price prediction in the existing technology is solved, and more accurate and real-time price prediction is achieved.
Patent Information
- Application Number
- CN202411933584.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-25
- Publication Date
- 2025-05-30
AI Technical Summary
In the rapidly changing market environment, the existing housing valuation model has lags in predicting prices and cannot effectively reflect market changes.
By obtaining the listing feature information of sample data, the property valuation model is optimized, so that its predicted price converges to the label price, and improves the accuracy of price prediction.
The optimized housing valuation model can predict prices more accurately, overcome the problem of market lag, and improve the real-time and accuracy of predictions.
Smart Images
Figure CN120069911A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of computer technologies, and in particular, to a method, device, medium, and computer program product for optimizing a housing price estimation model. Background Art
[0002] In the second-hand housing business, the second-hand housing price is a key factor determining whether the buyer and the seller can conclude a transaction. In the current market where the number of listed housing sources is surging, the ratio of the number of listed transactions to the number of listings is constantly increasing, and there are deviations in the market price perception between the buyer and the seller. To align the deviations in the market price perception between the buyer and the seller, it is necessary to give a price prediction result based on the current market.
[0003] Currently, the housing price estimation model usually predicts the second-hand housing price based on the information of the second-hand housing transaction housing sources. However, in a market where the housing price trend changes rapidly, the information of the second-hand housing transaction housing sources has a certain lag, resulting in a certain lag in the predicted price. Summary of the Invention
[0004] To solve the above technical problems or at least partially solve the above technical problems, the present disclosure provides a method, device, medium, and computer program product for optimizing a housing price estimation model.
[0005] An embodiment of the present disclosure provides a method for optimizing a housing price estimation model, the method including:
[0006] Obtaining sample data for optimizing the housing price estimation model, where the sample data includes the listing feature information of the sample housing sources and the first price of the sample housing sources;
[0007] Processing the listing feature information of the sample housing sources based on the housing price estimation model to obtain the second price of the sample housing sources;
[0008] Adjusting the model parameters of the housing price estimation model based on a preset loss function until the second price converges to the first price, to obtain an optimized housing price estimation model.
[0009] In some embodiments, the listing feature information includes:
[0010] At least one categorical feature and at least one numerical feature;
[0011] Wherein, the categorical feature is a feature related to the market; the numerical feature is a feature related to the housing source.
[0012] In some embodiments, the at least one categorical feature includes at least one of the following:
[0013] Location feature, first housing feature, first community feature;
[0014] Among them, the location characteristics include at least one of the following:
[0015] City code, urban area code, business district code, community code;
[0016] The first housing characteristics include at least one of the following:
[0017] Orientation category, housing type structure characteristics, building type;
[0018] The first community characteristics include:
[0019] Whether there is an elevator.
[0020] In some embodiments, the at least one numerical characteristic includes at least one of the following:
[0021] Second housing characteristics, second community characteristics;
[0022] Among them, the second housing characteristics include at least one of the following:
[0023] Number of bedrooms, number of living rooms, number of kitchens, number of bathrooms, number of balconies, area, whether it is full five, whether it is unique, whether it is a school district house, whether it is the top floor, whether it is the bottom floor, whether it is a high-rise building;
[0024] The second community characteristics include at least one of the following:
[0025] Number of floors, total number of floors, floor characteristics, community greening rate, plot ratio, number of underground parking spaces; building age characteristics; whether there is a fixed parking space; whether people and vehicles are separated; whether it is a closed community, decoration category, subway distance category.
[0026] In some embodiments, the preset loss function is determined based on a first loss sub-function and a second loss sub-function;
[0027] Among them, the first loss sub-function is a function independent of the listed price of the sample housing source; the second loss sub-function is a function related to the listed price of the sample housing source.
[0028] In some embodiments, the first loss sub-function is the mean absolute percentage error loss function;
[0029] And / or, the second loss sub-function is a loss function for the first business objective, and the first business objective is to limit the predicted price to be less than the listed price.
[0030] In some embodiments, the preset loss function is determined based on a first loss sub-function and a second loss sub-function, including:
[0031] The preset loss function is determined based on the weighted average of the first loss sub-function and the second loss sub-function.
[0032] In some embodiments, the weight of the first loss sub - function is:
[0033] The weighted average of the price differences corresponding to multiple groups of sample data;
[0034] Wherein, the price difference corresponding to each group of the sample data is determined based on the difference between the first price and the second price of the sample housing sources in the sample data;
[0035] The weight of the price difference corresponding to each group of the sample data is determined based on the first price of the sample housing sources in the sample data.
[0036] In some embodiments, the first loss sub - function is a loss function for a second business objective, and the second business objective is that the value of the mean absolute percentage error loss function is less than a preset value, and the preset value is a positive number less than 1.
[0037] Embodiments of the present disclosure also provide an electronic device, which includes: a processor; a memory for storing executable instructions that can be executed by the processor; the processor is configured to read the executable instructions from the memory and execute the instructions to implement the method for optimizing the housing price evaluation model provided by the embodiments of the present disclosure.
[0038] Embodiments of the present disclosure also provide a computer - readable storage medium, and the storage medium stores a computer program, and the computer program is used to execute the method for optimizing the housing price evaluation model provided by the embodiments of the present disclosure.
[0039] Embodiments of the present disclosure also provide a computer program product, including a computer program, and the computer program, when executed by a processor, implements the method for optimizing the housing price evaluation model provided by the embodiments of the present disclosure.
[0040] The above - mentioned technical solutions provided by the embodiments of the present disclosure, by obtaining sample data for optimizing the housing price evaluation model, can obtain the listing feature information of the sample housing sources, and the listing feature information is the information that can reflect market changes. Therefore, optimizing the housing price evaluation model based on the listing feature information of the sample housing sources enables the predicted price of the optimized housing price evaluation model to converge to the labeled price of the sample housing sources, improves the accuracy of price prediction, and overcomes the problem of market lag in the predicted price.
[0041] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understandable through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] The accompanying drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present disclosure and, together with the specification, are used to explain the principles of the present disclosure.
[0043] To more clearly illustrate the technical solutions in the embodiments of the present disclosure or in the prior art, the following will briefly introduce the accompanying drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.
[0044] Figure 1 It is a schematic flowchart of a method for optimizing a housing price evaluation model provided by an embodiment of the present disclosure;
[0045] Figure 2 It is a schematic diagram of a process for predicting the price of second-hand housing through training with different housing features using a deep neural network DNN model provided by an embodiment of the present disclosure;
[0046] Figure 3 It is a schematic diagram of a WDL model structure provided by an embodiment of the present disclosure;
[0047] Figure 4 It is a schematic diagram of an Autodis algorithm provided by an embodiment of the present disclosure;
[0048] Figure 5 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure. Detailed implementation manners
[0049] In order to be able to more clearly understand the above objects, features, and advantages of the present disclosure, the following will further describe the solutions of the present disclosure. It should be noted that, without conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other.
[0050] Many specific details are set forth in the following description in order to fully understand the present disclosure, but the present disclosure can also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only a part of the embodiments of the present disclosure, rather than all of the embodiments.
[0051] Figure 1 It is a schematic flowchart of a method for optimizing a housing price evaluation model provided by an embodiment of the present disclosure. This method can be executed by a housing price evaluation model optimization device, where the device can be implemented by software and / or hardware and is generally integrated in an electronic device. As Figure 1 shown, this method mainly includes the following steps 101 to 103:
[0052] 101. Obtain sample data for optimizing the housing price estimation model. The sample data includes the listing characteristics information of the sample housing units and the first price of the sample housing units.
[0053] Among them, the first price of the sample housing unit is the actual transaction price of the sample housing unit. That is to say, the first price of the sample housing unit is used as the label price for optimizing the housing price estimation model to evaluate whether the predicted price of the housing price estimation model is accurate. The closer the predicted price is to the label price, the more accurate the prediction of the housing price estimation model is.
[0054] In this embodiment, the obtained sample data can be multiple groups of sample data, and each group of sample data is used to train the housing price estimation model once. The multiple groups of sample data can train the housing price prediction model separately, or can jointly or simultaneously train the housing price prediction model.
[0055] 102. Process the listing characteristics information of the sample housing units based on the housing price estimation model to obtain the second price of the sample housing units.
[0056] Among them, the second price of the sample housing unit is the predicted price of the sample housing unit by the housing price estimation model. Therefore, there is a deviation between the second price of the sample housing unit and the first price of the sample housing unit. The purpose of optimizing the housing price estimation model is to make the predicted price (i.e., the second price) of the housing price estimation model close to or equal to the label price (i.e., the first price).
[0057] In this embodiment, the listing characteristics information of the sample housing unit is used as the input of the housing price estimation model. Input it into the housing price estimation model, and the housing price estimation model processes the listing characteristics information of the sample housing unit and outputs the second price of the sample housing unit.
[0058] 103. Adjust the model parameters of the housing price estimation model based on a preset loss function until the second price converges to the first price, and obtain the optimized housing price estimation model.
[0059] In this embodiment, based on the second price (i.e., the predicted price) of the sample housing unit obtained in step 102 and the first price (i.e., the label price) of the sample housing unit obtained in step 101, the value of the preset loss function can be calculated. Then, based on the value of the preset loss function, the model parameters of the housing price estimation model are adjusted through existing methods such as backpropagation until the second price converges to the first price, indicating that the optimization is completed, and the optimized housing price estimation model is obtained.
[0060] It can be seen that in this embodiment, by obtaining the sample data for optimizing the housing price evaluation model, the listing feature information of the sample housing can be obtained. The listing feature information is the information that can reflect the market changes. Therefore, based on the listing feature information of the sample housing, the housing price evaluation model is optimized, so that the predicted price (i.e., the second price) of the optimized housing price evaluation model converges to the labeled price (i.e., the first price) of the sample housing, improving the price prediction accuracy and overcoming the problem of market lag in the predicted price.
[0061] In some embodiments, the listing feature information includes:
[0062] At least one categorical feature and at least one numerical feature;
[0063] Among them, the categorical feature is a feature related to the market; the numerical feature is a feature related to the housing.
[0064] For example, the listing feature information includes 8 categorical features:
[0065] Location features: city code, urban area code, business district code, community code;
[0066] First housing features: orientation category, housing type structure feature, building type;
[0067] First community features: whether there is an elevator.
[0068] For another example, the listing feature information includes 24 numerical features:
[0069] Second housing features: number of bedrooms, number of living rooms, number of kitchens, number of bathrooms, number of balconies, area, whether it is more than five years old, whether it is the only one, whether it is a school district house, whether it is on the top floor, whether it is on the bottom floor, whether it is a high-rise building;
[0070] Second community features: number of floors, total number of floors, floor features, community greening rate, plot ratio, number of underground parking spaces; building age feature; whether there is a fixed parking space; whether it is a vehicle and pedestrian separation; whether it is a closed community, decoration category, subway distance category.
[0071] In some embodiments, in step 102, the housing price evaluation model can be a Neural Networks (NN) model. The network structure of the NN model usually consists of an input layer, a hidden layer, and an output layer. The housing price evaluation model can also be a predictive regression model. For the predictive regression model, the input layer receives the feature data, the hidden layer processes the data through an activation function, and the output layer generates continuous predicted values. For example, a simple fully connected neural network for housing price prediction may include multiple input layers, multiple hidden layers, each hidden layer using the ELU activation function, and finally an output layer using a linear activation function to predict continuous values. Figure 2Shows the process of training a Deep Neural Network (DNN) model using different housing source features to predict the price of second-hand housing sources.
[0072] In some embodiments, based on the basic DNN model, the Wide&Deep Learning (WDL) model is used to optimize the memory ability of the existing housing source valuation model for artificial cross features, and further combines the Autodis algorithm to optimize the discretized feature representation of continuous features.
[0073] Among them, the WDL model structure is as Figure 3 shown. In Figure 3 , the wide part is a generalized linear model, and the input features are the original input features and transformed features. The deep part is a feedforward neural network. Sparse and high-dimensional categorical features are transformed into low-dimensional dense vector learning vector parameters. The wide part and the deep part are jointly trained and optimized simultaneously.
[0074] The Autodis algorithm is as Figure 4 shown. In Figure 4 : (1) Defines a set of Meta-Embeddings for the continuous features in each domain.
[0075] (2) Adopts a two-layer neural network, inputs the continuous feature values, and outputs the continuous feature values assigned to different Meta-Embeddings buckets.
[0076] (3) Aggregates the embedding results of multiple buckets to obtain the final embedding of the continuous feature value. Among them, the aggregation method can be any one of the following A to C:
[0077] A. Max pooling: Selects the Meta-Embeddings bucket with the highest probability value as the final embedding representation of the feature value.
[0078] B. Top-K summation: Selects the top-k embeddings with the highest probability and sums them.
[0079] C. Weighted average: Weighted average with probability logical values.
[0080] In some embodiments, in step 103, the preset loss function is determined based on the first loss sub-function and the second loss sub-function. Among them, the first loss sub-function is a function independent of the listed price of the sample housing source; the second loss sub-function is a function related to the listed price of the sample housing source. Among them, the second loss sub-function introduces the listed price, which can strengthen the housing source valuation model's perception of the current market trend.
[0081] Embodiment 1
[0082] The first loss sub - function is the Mean Absolute Percentage Error (MAPE) loss function. And / or, the second loss sub - function is a loss function for the first business objective, where the first business objective is to limit the predicted price to be less than the listed price.
[0083] The preset loss function is determined based on the first loss sub - function and the second loss sub - function, including: the preset loss function is determined based on the weighted average of the first loss sub - function and the second loss sub - function.
[0084] Loss = α·Loss1 + β·Loss2
[0085]
[0086] Loss2 = max(0, y p -listing - y)
[0087]
[0088] Among them, Loss1 is the MAPE loss function (i.e., the first loss sub - function), Loss2 is the two - segment listing loss function (i.e., the second loss sub - function), and α and β are the weights of Loss1 and Loss2 respectively. y t is the first price (i.e., the label price), y p is the second price (i.e., the predicted price). ∈ is a preset constant used to avoid the situation of division by zero. For example, ∈ is a number close to zero but greater than zero. Listing_y is the listed price. Final Loss is the preset loss function. n is the number of groups of sample data.
[0089] Example 2
[0090] The weight α of the first loss sub - function is: the weighted average of the price differences corresponding to multiple groups of sample data. Among them, the price difference corresponding to each group of sample data is determined based on the difference between the first price (i.e., the label price) and the second price (i.e., the predicted price) of the sample housing sources in the sample data. The weight of the price difference corresponding to each group of sample data is determined based on the first price of the sample housing sources in the sample data.
[0091] For example, to make the housing price evaluation model have a bias towards the prediction accuracy of high - price housing sources, in this embodiment, based on the first price (i.e., the label price), a normalized weight design for high - price housing source weights is carried out on the basis of the first loss sub - function, that is, the MAPE loss function, and this optimization direction is followed during the training process. The expression of the weight α of the first loss sub - function is as follows:
[0092]
[0093] Among them, WeightMAPE is the weight α of the first loss sub - function, n is the number of groups of sample data. t represents the t - th group of sample data. ω represents the weight calculated for each group of sample data according to the first price (i.e., the label price) in batch training, that is, the weight of the price difference corresponding to each group of sample data.
[0094] Among them, is the price difference corresponding to the sample data, y t represents the first price (i.e., the label price) of the sample housing unit, y p represents the second price (i.e., the predicted price) of the sample housing unit, ∈ is a preset constant used to avoid the situation of division by zero. For example, ∈ is a number close to zero but greater than zero.
[0095] Example 3
[0096] The first loss sub - function is the loss function for the second business objective. The second business objective is that the value of the mean absolute percentage error loss function is less than a preset value, and the preset value is a positive number less than 1.
[0097] For example, the second business objective is that the MAPE loss function is less than 3%. The MAPE loss function is designed as follows:
[0098]
[0099] Among them, y t,i represents the first price (i.e., the label price) of the sample housing unit in the i - th group of sample data, y p,i represents the second price (i.e., the predicted price) of the sample housing unit in the i - th group of sample data. ∈ is a preset constant used to avoid the situation of division by zero. For example, ∈ is a number close to zero but greater than zero.
[0100] The design of this loss function aims to make the learning of the model weights more focused on optimizing those housing units whose predicted values and target values do not reach an error of less than 3% in the current stage.
[0101] Corresponding to the foregoing housing unit valuation model optimization method, an embodiment of the present disclosure further provides a housing unit valuation model optimization device. The device can be implemented by software and / or hardware and is generally integrated in an electronic device. The housing unit valuation model optimization device includes:
[0102] An acquisition unit, configured to acquire sample data for optimizing the housing unit valuation model. The sample data includes the listing feature information of the sample housing unit and the first price of the sample housing unit;
[0103] A processing unit, configured to process the listing feature information of the sample housing units based on a housing price estimation model to obtain the second price of the sample housing units;
[0104] An adjustment unit, configured to adjust the model parameters of the housing price estimation model based on a preset loss function until the second price converges to the first price, so as to obtain an optimized housing price estimation model.
[0105] In some embodiments, the listing feature information includes:
[0106] At least one categorical feature and at least one numerical feature;
[0107] Wherein, the categorical feature is a feature related to the market; the numerical feature is a feature related to the housing unit.
[0108] In some embodiments, the at least one categorical feature includes at least one of the following:
[0109] Location feature, first housing feature, first community feature;
[0110] Wherein, the location feature includes at least one of the following:
[0111] City code, urban area code, business district code, community code;
[0112] The first housing feature includes at least one of the following:
[0113] Orientation category, housing type structure feature, building type;
[0114] The first community feature includes:
[0115] Whether there is an elevator.
[0116] In some embodiments, the at least one numerical feature includes at least one of the following:
[0117] Second housing feature, second community feature;
[0118] Wherein, the second housing feature includes at least one of the following:
[0119] Number of bedrooms, number of living rooms, number of kitchens, number of bathrooms, number of balconies, area, whether it is full five years, whether it is the only one, whether it is a school district house, whether it is on the top floor, whether it is on the bottom floor, whether it is a high-rise building;
[0120] The second community feature includes at least one of the following:
[0121] Number of floors, total number of floors, floor feature, community greening rate, plot ratio, number of underground parking spaces; building age feature; whether there is a fixed parking space; whether it is a vehicle and pedestrian separation; whether it is a closed community, decoration category, subway distance category.
[0122] In some embodiments, the preset loss function is determined based on a first loss sub-function and a second loss sub-function;
[0123] wherein, the first loss sub-function is a function independent of the listing price of the sample housing unit; the second loss sub-function is a function related to the listing price of the sample housing unit.
[0124] In some embodiments, the first loss sub-function is the mean absolute percentage error loss function;
[0125] and / or, the second loss sub-function is a loss function for a first business objective, and the first business objective is to limit the predicted price to be less than the listing price.
[0126] In some embodiments, determining the preset loss function based on the first loss sub-function and the second loss sub-function includes:
[0127] The preset loss function is determined based on the weighted average of the first loss sub-function and the second loss sub-function.
[0128] In some embodiments, the weight of the first loss sub-function is:
[0129] The weighted average of the price differences corresponding to multiple groups of sample data;
[0130] wherein, the price difference corresponding to each group of sample data is determined based on the difference between the first price and the second price of the sample housing unit in the sample data;
[0131] The weight of the price difference corresponding to each group of sample data is determined based on the first price of the sample housing unit in the sample data.
[0132] In some embodiments, the first loss sub-function is a loss function for a second business objective, and the second business objective is that the value of the mean absolute percentage error loss function is less than a preset value, and the preset value is a positive number less than 1.
[0133] The housing unit valuation model optimization device provided by the embodiments of the present disclosure can execute the housing unit valuation model optimization method provided by any embodiment of the present disclosure, and has corresponding functional modules and beneficial effects for executing the method.
[0134] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working process of the device embodiments described above can refer to the corresponding process in the method embodiments, and will not be elaborated herein.
[0135] The embodiments of the present disclosure provide an electronic device, and the electronic device includes: a storage device on which a computer program is stored; a processing device for executing the computer program in the storage device to implement the steps of any method in the present disclosure.
[0136] Next, refer toFigure 5 , which shows a schematic structural diagram of an electronic device 500 suitable for implementing the embodiments of the present disclosure. The terminal devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 5 The electronic device shown is only an example and should not impose any limitations on the functions and usage scope of the embodiments of the present disclosure.
[0137] As Figure 5 shown, the electronic device 500 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 501, which may perform various appropriate actions and processes according to the programs stored in the read-only memory (ROM) 502 or the programs loaded from the storage device 508 into the random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. The input / output (I / O) interface 505 is also connected to the bus 504.
[0138] Generally, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509. The communication device 509 may allow the electronic device 500 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 5 the electronic device 500 with various devices is shown, it should be understood that it is not required to implement or have all the shown devices. More or fewer devices may be alternatively implemented or had.
[0139] Particularly, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, the embodiments of the present disclosure include a computer program carried on a non-transitory computer-readable medium, and the computer program includes program codes for performing the methods shown in the flowcharts. In such an embodiment, the computer program may be downloaded and installed from the network through the communication device 509, or installed from the storage device 508, or installed from the ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the methods of the embodiments of the present disclosure are executed.
[0140] In addition to the above methods and devices, embodiments of the present disclosure may also be computer program products, which include computer program instructions that, when run by a processor, cause the processor to execute the methods provided by the embodiments of the present disclosure. The computer program products may be written in any combination of one or more programming languages for programming code to perform the operations of the embodiments of the present disclosure. The programming languages include object-oriented programming languages such as Java, C++, etc., and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's device, executed as a stand-alone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0141] In addition, embodiments of the present disclosure may also be computer-readable storage media, on which computer program instructions are stored, and the computer program instructions, when run by a processor, cause the processor to execute the housing price evaluation model optimization method provided by the embodiments of the present disclosure.
[0142] The computer-readable storage media may adopt any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may, for example, include but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or components, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: electrical connections with one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.
[0143] Embodiments of the present disclosure also provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the housing price evaluation model optimization method in the embodiments of the present disclosure.
[0144] It can be understood that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.
[0145] For example, when responding to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested by the user will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, an application program, a server, or a storage medium that performs the operations of the present disclosed technical solution according to the prompt message.
[0146] As an optional but non-limiting implementation manner, the manner of sending a prompt message to the user in response to receiving an active request from the user may, for example, be in the form of a pop-up window, and the prompt message may be presented in text in the pop-up window. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.
[0147] It can be understood that the above process of notifying and obtaining user authorization is only illustrative and does not constitute a limitation on the implementation manner of the present disclosure. Other manners that comply with relevant laws and regulations can also be applied to the implementation manner of the present disclosure.
[0148] It should be noted that in this article, 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 any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising a..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.
[0149] The above are only specific implementation manners of the present disclosure, enabling those skilled in the art to understand or implement the present disclosure. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can 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 these embodiments described herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.
Claims
1. A housing valuation model optimization method, characterized in that: include: Acquire sample data for optimizing a house valuation model, wherein the sample data includes listing feature information of a sample house and a first price of the sample house; Processing the listing feature information of the sample house based on the house valuation model to obtain a second price of the sample house; The model parameters of the housing valuation model are adjusted based on a preset loss function until the second price converges to the first price, thereby obtaining an optimized housing valuation model.
2. The method according to claim 1, characterized in that: The listing feature information includes: at least one categorical feature and at least one numerical feature; Among them, the category features are features related to the market; and the numerical features are features related to housing sources.
3. The method according to claim 2, characterized in that The at least one category feature includes at least one of the following: Locational characteristics, primary housing characteristics, primary neighborhood characteristics; The location characteristics include at least one of the following: City code, urban district code, business district code, and community code; The first housing feature includes at least one of the following: Orientation category, apartment structure characteristics, and building type; The first cell characteristics include: Is there an elevator? 4. The method according to claim 2, characterized in that: The at least one numerical feature includes at least one of the following: Second house characteristics, second neighborhood characteristics; Wherein, the second housing feature includes at least one of the following: Number of bedrooms, living rooms, kitchens, bathrooms, balconies, area, whether it is full of five, whether it is the only one, whether it is a school district house, whether it is the top floor, whether it is the ground floor, whether it is a high-rise building; The second cell characteristic includes at least one of the following: Number of floors, total number of floors, floor characteristics, community greening rate, volume ratio, number of underground parking spaces; building age characteristics; whether there are fixed parking spaces; whether there are separations between people and vehicles; whether the community is closed, decoration type, and subway distance type.
5. The method according to claim 1, characterized in that The preset loss function is determined based on the first loss sub-function and the second loss sub-function; The first loss sub-function is a function that is independent of the listing price of the sample house; and the second loss sub-function is a function that is related to the listing price of the sample house.
6. The method according to claim 5, characterized in that The first loss sub-function is a mean absolute percentage error loss function; And / or, the second loss sub-function is a loss function for a first business objective, and the first business objective is to limit the predicted price to be less than the listed price.
7. The method according to claim 5 or 6, characterized in that: The preset loss function is determined based on the first loss sub-function and the second loss sub-function, and includes: The preset loss function is determined based on a weighted average of the first loss sub-function and the second loss sub-function.
8. The method according to claim 7, characterized in that The weight of the first loss sub-function is: The weighted average of the price differences corresponding to multiple groups of sample data; The price difference corresponding to each group of the sample data is determined based on the difference between the first price of the sample house in the sample data and the second price of the sample house; The weight of the price difference corresponding to each group of the sample data is determined based on the first price of the sample housing in the sample data.
9. The method according to claim 5 or 6, characterized in that: The first loss sub-function is a loss function for a second business objective, and the second business objective is that the value of the mean absolute percentage error loss function is less than a preset value, and the preset value is a positive number less than 1.
10. An electronic device, characterized in that: The electronic device comprises: a storage device having a computer program stored thereon; A processing device is used to execute the computer program in the storage device to implement the steps of the property valuation model optimization method described in any one of claims 1-9.
11. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and the computer program is used to execute the property valuation model optimization method described in any one of claims 1-9 above.
12. A computer program product, characterized in that It includes a computer program, which, when executed by a processor, implements the housing valuation model optimization method described in any one of claims 1 to 9.