CN115345684A - House rent estimation method and device, terminal equipment and readable storage medium - Google Patents
House rent estimation method and device, terminal equipment and readable storage medium Download PDFInfo
- Publication number
- CN115345684A CN115345684A CN202211269958.1A CN202211269958A CN115345684A CN 115345684 A CN115345684 A CN 115345684A CN 202211269958 A CN202211269958 A CN 202211269958A CN 115345684 A CN115345684 A CN 115345684A
- Authority
- CN
- China
- Prior art keywords
- house
- neural network
- rent
- preset
- information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000013528 artificial neural network Methods 0.000 claims abstract description 132
- 239000013598 vector Substances 0.000 claims abstract description 95
- 238000012549 training Methods 0.000 claims description 131
- 230000000306 recurrent effect Effects 0.000 claims description 62
- 238000013527 convolutional neural network Methods 0.000 claims description 39
- 230000006870 function Effects 0.000 claims description 21
- 238000010801 machine learning Methods 0.000 abstract description 4
- 230000008569 process Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000013277 forecasting method Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0278—Product appraisal
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0645—Rental transactions; Leasing transactions
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a device for pre-estimating house rent, terminal equipment and a readable storage medium, belonging to the field of machine learning, wherein the method for pre-estimating the house rent comprises the following steps: acquiring house information, wherein the house information comprises a house structure plan and house structural information; and extracting a first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting a pre-estimated rent. The house rent is estimated by inputting the house structural information and the house structure plane graph into the fully-connected regression neural network trained in advance, the problem that consideration factors caused by manually estimating rent are not comprehensive and objective enough is solved, the purpose of estimating house rent more comprehensively, accurately and objectively is achieved, and the accuracy and the efficiency of estimating the house rent in the market are improved.
Description
Technical Field
The invention relates to the field of machine learning, in particular to a house rent estimation method, a house rent estimation device, terminal equipment and a readable storage medium.
Background
The house rent is influenced by a plurality of factors, the factors of each set of house are different, and the reasonable rent of one house is unreasonable through artificial estimation.
Some existing schemes evaluate house rent in a quantitative manner, and specifically, return the house rent by using structured data and using a machine learning model. However, for whether the unstructured data such as the structure in the suite is reasonable or not, whether the unstructured data is reasonable or not is judged manually, the unstructured data is converted into structured data and then further processed, a certain amount of objective information is lost, and therefore the evaluation result is not comprehensive and accurate enough.
Therefore, how to comprehensively, accurately and objectively estimate house rent is a problem to be solved urgently at present.
Disclosure of Invention
The application mainly aims to provide a house rent estimation method, a house rent estimation device, terminal equipment and a readable storage medium, and aims to solve the problem of comprehensively, accurately and objectively estimating house rent.
In order to achieve the purpose, the application provides a house rent estimation method, which is applied to the technical field of machine learning and comprises the following steps:
acquiring house information, wherein the house information comprises a house structure plane graph and house structural information;
and extracting a first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting a pre-estimated rent.
Optionally, the step of extracting the first feature vector from the house information and inputting the first feature vector into a preset fully-connected recurrent neural network, and outputting the pre-estimated rent further includes:
acquiring historical house information, wherein the historical house information comprises a historical house structure plan and historical house structure information;
extracting a second feature vector from the house historical information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network;
inputting the second feature vector into the fully-connected recurrent neural network, and outputting the estimated rent of the house;
updating the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network according to the estimated rent and the historical rent in the house historical information, a preset loss function and a preset learning rate;
and executing the step of acquiring the house historical information according to the preset training times, wherein the house historical information comprises a house structure historical plan and house structural historical information until the preset training times are completed.
Optionally, the step of extracting the second feature vector from the house history information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network includes:
extracting a third feature vector from the building structure historical plan through the convolutional neural network VGG16 pre-training model;
inputting the house structural historical information into the fully-connected neural network, and outputting a corresponding fourth feature vector;
and fusing the third feature vector and the fourth feature vector to obtain a fused second feature vector.
Optionally, the step of updating the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model, and the weight of the fully-connected neural network according to the estimated rent and the historical rent in the historical house information, a preset loss function, and a preset learning rate includes:
making a difference between the estimated rent and the rent in the house historical information, and calculating loss;
calculating a gradient according to the loss and the preset loss function to obtain a gradient descending direction;
and updating the weight of the fully-connected recurrent neural network, the weight of the pre-training model of the convolutional neural network VGG16 and the weight of the fully-connected neural network according to the gradient descent direction and the preset learning rate.
Optionally, the obtaining of the house historical information is performed according to preset training times, where the house historical information includes a house structure historical plan and house structural historical information, and the step of completing the preset training times includes:
accumulating the training times and comparing the training times with preset training times;
if the preset training times are reached, stopping training the fully-connected recurrent neural network;
and if the preset training times are not reached, executing the step of acquiring the house historical information until the training times reach the preset training times.
Optionally, before the step of extracting the second feature vector from the house history information by using the preset convolutional neural network VGG16 pre-training model or the preset fully-connected neural network, the method further includes:
initializing weights of the fully-connected neural network and the fully-connected recurrent neural network;
and setting the learning rate and the training times of the fully-connected recurrent neural network.
Optionally, the step of obtaining the house history information includes:
and randomly acquiring the house historical information according to normal distribution.
The application embodiment also provides a device for pre-estimating the house rent, and the device for pre-estimating the house rent comprises:
the acquisition module is used for acquiring house information;
and the estimation module is used for extracting the first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting an estimated rent.
The embodiment of the application also provides a terminal device, which comprises a memory, a processor and a house rent estimation program which is stored on the memory and can run on the processor, wherein the house rent estimation program is executed by the processor to realize the house rent estimation method.
The implementation case of the application also provides a readable storage medium, wherein a house rent estimation program is stored on the readable storage medium, and the house rent estimation program is executed by a processor to realize the house rent estimation method.
The application provides an analysis method and device of component dependency relationship, a terminal device and a readable storage medium. The method comprises the steps that house information is obtained, wherein the house information comprises a house structure plane graph and house structural information; and extracting a first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting a pre-estimated rent. Based on the method and the device, the house structural information and the house structure plane graph are input into the fully-connected regression neural network trained in advance, so that house rent is estimated, the problem that consideration factors caused by artificial estimation of rent are not comprehensive and objective enough is solved, the purpose of estimating house rent more comprehensively, accurately and objectively is achieved, and the accuracy and the efficiency of estimation of house rent in the market are improved.
Drawings
FIG. 1 is a schematic diagram of a functional module of a terminal to which the method for estimating house rent belongs;
FIG. 2 is a schematic flow chart of a first exemplary embodiment of a method for forecasting house rent according to the present application;
FIG. 3 is a schematic flow chart of a second exemplary embodiment of a house rent estimation method according to the present application;
fig. 4 is a schematic flowchart of a third exemplary embodiment of a house rent estimation method according to the present application;
FIG. 5 is a schematic flow chart illustrating a fourth exemplary embodiment of a house rent forecasting method according to the present application;
FIG. 6 is a schematic flow chart of a fifth exemplary embodiment of a house rent estimation method according to the present application;
FIG. 7 is a schematic flow chart of a sixth exemplary embodiment of a house rent forecasting method according to the present application;
fig. 8 is a schematic flowchart of a seventh exemplary embodiment of a house rent estimation method according to the present application;
FIG. 9 is a block diagram of the structural process of the single training fully-connected recurrent neural network of the present application.
The implementation, functional features and advantages of the objectives of the present application will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The method comprises the steps of obtaining house information, wherein the house information comprises a house structure plan and house structural information; and extracting a first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting the estimated rent. Based on the scheme, the house structural information and the house structure plane graph are input into the fully-connected recurrent neural network trained in advance, so that house rent is estimated, the problem that consideration factors caused by artificial estimation of rent are not comprehensive and objective enough is solved, the purpose of estimating house rent more comprehensively, accurately and objectively is achieved, and the accuracy and the efficiency of estimation of house rent in the market are improved.
Specifically, referring to fig. 1, fig. 1 is a schematic diagram of functional modules of a terminal device to which the estimation apparatus for house rent belongs. The house rent estimation device is a terminal-based device capable of training a fully-connected recurrent neural network for estimating house rent to estimate the house rent, so that the purpose of estimating the house rent comprehensively, accurately and objectively is achieved, and the house rent estimation device can be borne on the terminal device in a hardware or software mode.
In this embodiment, the terminal device to which the forecast apparatus of the house rental belongs at least includes an output module 110, a processor 120, a memory 130 and a communication module 140.
The memory 130 stores an operating system and a house rent estimation program, and the house rent estimation device can acquire house information, wherein the house information comprises a house structure plan and house structure information; extracting a first feature vector from the house information, inputting the first feature vector into a preset full-connection regression neural network, outputting information such as estimated rent and the like, and storing the information in the memory 130; the output module 110 may be a display screen or the like. The communication module 140 may include a WIFI module, a mobile communication module, a bluetooth module, and the like, and communicates with an external device or a server through the communication module 140.
Wherein, the pre-estimation program of house rent in the memory 130, when executed by the processor, implements the following steps:
acquiring house information, wherein the house information comprises a house structure plan and house structural information;
and extracting a first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting a pre-estimated rent.
Further, the estimation procedure of house rent in the memory 130 when executed by the processor further realizes the following steps:
acquiring historical house information, wherein the historical house information comprises a historical house structure plan and historical house structure information;
extracting a second feature vector from the house historical information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network;
inputting the second feature vector into the fully-connected recurrent neural network, and outputting the estimated rent of the house;
updating the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network according to the estimated rent and the historical rent in the house historical information, a preset loss function and a preset learning rate;
and executing the step of acquiring the house historical information according to the preset training times, wherein the house historical information comprises a house structure historical plan and house structural historical information until the preset training times are completed.
Further, the house rent estimator in the memory 130, when executed by the processor, further performs the steps of:
extracting a third feature vector from the building structure historical plan through the convolutional neural network VGG16 pre-training model;
inputting the house structural historical information into the fully-connected neural network, and outputting a corresponding fourth feature vector;
and fusing the third feature vector and the fourth feature vector to obtain a fused second feature vector.
Further, the estimation procedure of house rent in the memory 130 when executed by the processor further realizes the following steps:
making a difference between the estimated rent and the rent in the house historical information, and calculating loss;
calculating a gradient according to the loss and the preset loss function to obtain a gradient descending direction;
and updating the weight of the fully-connected regression neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network according to the gradient descending direction and the preset learning rate.
Further, the house rent estimator in the memory 130, when executed by the processor, further performs the steps of:
accumulating the training times and comparing the training times with the preset training times;
if the preset training times are reached, stopping training the fully-connected recurrent neural network;
and if the preset training times are not reached, executing the step of acquiring the house historical information until the training times reach the preset training times.
Further, the house rent estimator in the memory 130, when executed by the processor, further performs the steps of:
initializing weights of the fully-connected neural network and the fully-connected recurrent neural network;
and setting the learning rate and the training times of the fully-connected recurrent neural network.
Further, the estimation procedure of house rent in the memory 130 when executed by the processor further realizes the following steps:
and randomly acquiring the house historical information according to normal distribution.
Based on the above terminal device architecture but not limited to the above architecture, embodiments of the method of the present application are provided.
Referring to fig. 2, fig. 2 is a schematic flow chart of a house rent forecasting method according to a first exemplary embodiment. The house rent estimation method comprises the following steps:
step S110, house information is obtained, and the house information comprises a house structure plane graph and house structural information;
specifically, the acquired house information is used as a data basis for estimating rent, and specifically comprises a house structure plan and house structural information, the house structural information specifically comprises information such as processed and structured areas, building ages, floors, elevator existence, distance from a subway station and the like, the processing mode comprises dimension removal and invalid data removal, and the data is divided into parts according to the corresponding relation and by taking a house as a unit to be structured.
And step S130, extracting a first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting a pre-estimated rent.
Specifically, the preset fully-connected recurrent neural network is obtained by training house historical information, and the specific training mode is to initialize the weights of the fully-connected neural network and the fully-connected recurrent neural network; and setting the learning rate and the training times of the fully-connected recurrent neural network. Acquiring historical house information, wherein the historical house information comprises a historical house structure plan and historical house structural information; extracting a second feature vector from the house historical information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network; inputting the second feature vector into the fully-connected recurrent neural network, and outputting the estimated rent of the house; updating the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network according to the estimated rent and the historical rent in the historical information of the house, a preset loss function and a preset learning rate; accumulating the training times and comparing the training times with preset training times; if the preset training times are reached, stopping training the fully-connected recurrent neural network; if the preset training times are not reached, executing the step of acquiring the house historical information until the training times reach the preset training times, obtaining a fully-connected regression neural network capable of accurately and comprehensively estimating house rent after training is completed, and finally acquiring the house information in the step S110; and fusing the third feature vector and the fourth feature vector to obtain a fused second feature vector, inputting the fused second feature vector to a trained fully-connected recurrent neural network, and outputting the estimated rent.
According to the scheme, the house information is obtained specifically, and the house information comprises a house structure plan and house structural information; and extracting a first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting a pre-estimated rent. Based on the scheme, the house rent can be estimated according to the house structural information and the house structure plane graph by training the fully-connected recurrent neural network capable of estimating the house rent, so that the purpose of comprehensively, accurately and objectively estimating the house rent is achieved, and the accuracy and the efficiency of estimating the house rent in the market are improved.
Further, referring to fig. 3, fig. 3 is a schematic flowchart of a house rent estimation method according to a second exemplary embodiment, where the step of extracting a first feature vector from the house information and inputting the extracted first feature vector into a preset fully-connected recurrent neural network, and outputting the estimated rent further includes:
step S115, house historical information is obtained, and the house historical information comprises a house structure historical plan and house structural historical information;
specifically, the acquired house historical information is used as data for training a preset full-connection regression neural network, and specifically comprises a house structure historical plan and house structural historical information, the house structural historical information specifically comprises information such as processed and structured areas, historical rent, building age, floors, presence or absence of elevators, distance from a subway station and the like, the processing mode comprises dimension removal and invalid data removal, and the data are divided into parts according to corresponding relations and by taking a house as a unit for structuring.
Step S116, extracting a second feature vector from the house historical information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network;
inputting the house structural history information into the fully-connected neural network, and outputting a corresponding fourth feature vector, wherein the preset fully-connected neural network is only used for feature extraction; and fusing the third feature vector and the fourth feature vector to obtain a fused second feature vector.
Step S117, inputting the second feature vector into the fully-connected recurrent neural network, and outputting the estimated rent of the house;
specifically, the feature vectors are input as parameters for training the fully-connected recurrent neural network, then are accumulated through weights and are propagated forwards through a bias value and an activation function, and finally estimated rent is output through an output layer.
And step S118, updating the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network according to the estimated rent and the historical rent in the house historical information, a preset loss function and a preset learning rate.
Specifically, the estimated rent and the historical rent in the historical information are substituted into a loss function for calculation to obtain a gradient descending direction, and then the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network are updated through a backward propagation algorithm in combination with a preset learning rate.
Step S119, the step of obtaining the house historical information is executed according to the preset training times, and the house historical information comprises the steps of house structure historical plane graph and house structural historical information until the preset training times are completed.
Specifically, the training times are accumulated and compared with the preset training times; if the preset training times are reached, stopping training the fully-connected recurrent neural network; and if the preset training times are not reached, executing the step of acquiring the house historical information until the training times reach the preset training times.
According to the scheme, the historical house information is obtained, and the historical house information comprises a historical house structure plan and historical house structure information; extracting a second feature vector from the house historical information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network; inputting the second feature vector into the fully-connected recurrent neural network, and outputting the estimated rent of the house; updating the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network according to the estimated rent and the historical rent in the house historical information, a preset loss function and a preset learning rate; and executing the step of acquiring the house historical information according to the preset training times, wherein the house historical information comprises a house structure historical plan and house structural historical information until the preset training times are completed. Based on the scheme, the characteristic vector is extracted and input into the fully-connected recurrent neural network, the weight is adjusted through multiple times of training, the fully-connected recurrent neural network capable of estimating rent is obtained through training, and technical support is provided for subsequent comprehensive, accurate and objective estimation of house rent.
Further, referring to fig. 4, fig. 4 is a flowchart illustrating a house rent estimation method according to a third exemplary embodiment, where the step of extracting a second feature vector from the house history information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network includes:
step S1161, extracting a third feature vector from the house structure historical plan by the convolutional neural network VGG16 pre-training model;
specifically, the convolutional neural network VGG16 is composed of 13 convolutional layers and 3 full-connection layers, a picture with a certain size passes through the convolutional neural network VGG16 to obtain a multidimensional third feature vector, and the specific dimension is determined according to the size of the picture and the calculation resource.
Step S1162, inputting the house structuralization historical information into the fully-connected neural network, and outputting a corresponding fourth feature vector;
specifically, the house structured history information specifically includes information dimension of the processed and structured area, historical rent, building age, floor, presence or absence of elevator, distance from a subway station and the like, invalid data is removed, and the data is divided according to the corresponding relationship and by taking the house as a unit, so that structured data is obtained; and inputting the house structural historical information into the fully-connected neural network, and extracting a multidimensional fourth feature vector, wherein the specific dimension is determined according to the size of the image and the computing resource.
And step S1163, fusing the third feature vector and the fourth feature vector to obtain a fused second feature vector.
Specifically, the fusion condition of the third feature vector and the fourth feature vector is that the dimensions are the same, and a second feature vector is obtained after fusion so as to be used for training the fully-connected recurrent neural network.
In this embodiment, through the above scheme, a third feature vector is extracted from the house structure history plan by specifically using the convolutional neural network VGG16 pre-training model; and fusing the third feature vector and the fourth feature vector to obtain a fused second feature vector. Based on the scheme, the feature vectors are extracted and fused through the convolutional neural network VGG16 pre-training model and the fully-connected neural network respectively, and data support is provided for subsequent training of the feature vectors.
Further, referring to fig. 5, fig. 5 is a flowchart illustrating a house rent estimation method according to a fourth exemplary embodiment, where the updating of the weights of the fully-connected recurrent neural network, the convolutional neural network VGG16 pre-training model, and the fully-connected neural network according to the estimated rent and the historical rent in the house historical information, a preset loss function, and a preset learning rate includes:
step S1181, making a difference between the estimated rent and the rent in the house historical information, and calculating loss;
step S1182, calculating a gradient according to the loss and the preset loss function to obtain a gradient descending direction;
specifically, the partial derivative of the preset loss function is obtained through judgment, and the gradient descending direction is obtained according to the descending direction of the loss.
And step S1183, updating the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model, the weight of the fully-connected neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network according to the gradient descending direction and the preset learning rate.
Specifically, the gradient descent direction and the preset learning rate jointly control the size of the weight, so as to ensure that the problem that the prediction point with the minimum loss cannot be quickly found due to too large or too small weight change is solved, and the weight of each layer of the fully-connected recurrent neural network, the convolutional neural network VGG16 pre-training model and the fully-connected neural network is updated through a backward propagation algorithm.
According to the scheme, the loss is calculated by subtracting the estimated rent from the rent in the house historical information; calculating a gradient according to the loss and the preset loss function to obtain a gradient descending direction; and updating the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model, the weight of the fully-connected neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network according to the gradient descending direction and the preset learning rate. Based on the scheme, the weight is updated through a back propagation algorithm by calculating the gradient descending direction, the trained and optimized fully-connected recurrent neural network is obtained, and the accuracy of the prediction result of the fully-connected recurrent neural network is improved.
Further, referring to fig. 6, fig. 6 is a flowchart of a fifth exemplary embodiment of a house rent estimation method, where the acquiring house history information according to a preset number of training is executed, the house history information includes a house structure history plan and house structure history information, and the step of completing the preset number of training includes:
step S1191, add up the training times and compare with the preset training times;
step S1192, if the preset training times is reached, stopping training the fully-connected recurrent neural network;
step S1193, if the preset number of times of training is not reached, execute the step of obtaining the house history information until the number of times of training reaches the preset number of times of training.
According to the scheme, the training times are accumulated and compared with the preset training times; if the preset training times are reached, stopping training the fully-connected recurrent neural network; and if the preset training times are not reached, executing the step of acquiring the house historical information until the training times reach the preset training times. Based on the scheme, the training times of the fully-connected recurrent neural network are judged, so that the aim of determining whether to stop training is fulfilled, and the problem of resource waste caused by unlimited training is solved.
Further, referring to fig. 7, fig. 7 is a flowchart illustrating a fifth exemplary embodiment of a house rent estimation method, where before the step of extracting a second feature vector from the house history information by using a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network, the method further includes:
step S113, initializing the weight of the fully-connected neural network and the fully-connected recurrent neural network;
specifically, before the fully-connected neural network and the fully-connected recurrent neural network are used, the weights need to be initialized, so that the first calculation is convenient to use.
And step S114, setting the learning rate and the training times of the fully-connected recurrent neural network.
Specifically, the learning rate of the fully-connected recurrent neural network determines whether and when the objective function converges to a local minimum. An appropriate learning rate enables the objective function to converge to a local minimum in an appropriate time. And guiding us how to use the gradient of the loss function to adjust the hyperparameter of the network weight in the gradient descent method.
According to the scheme, the weight of the fully-connected neural network and the weight of the fully-connected recurrent neural network are initialized; and setting the learning rate and the training times of the fully-connected recurrent neural network. Based on the scheme, preparation work is made for subsequent use of the neural network, and the problem of resource waste can be avoided through appropriate learning rate setting and training times.
Further, referring to fig. 8, fig. 8 is a flowchart illustrating a house rent estimation method according to a sixth exemplary embodiment, where the step of obtaining house history information includes:
and step S1151, randomly acquiring the house history information according to normal distribution.
Specifically, the house history information is a group of extracted feature vectors and is input to the neural network, data closer to the actual situation is provided for better training the neural network, and the house history information is randomly acquired through normal distribution.
According to the embodiment, through the scheme, the house historical information is obtained randomly according to the normal distribution, the authenticity of training data is improved, and the accuracy of the result obtained by the neural network trained through the information in actual use is guaranteed.
Further, referring to fig. 9, fig. 9 is a block diagram of a structural flow of a single training fully-connected recurrent neural network. The specific process comprises the following steps:
building structure plan and building structure information (namely, the structure factors in FIG. 9); inputting the house structural information into a fully-connected neural network (namely a feature extraction network (multi layer Perceptron) MLP (multi layer Perceptron) in the figure 9), and outputting a corresponding fourth feature vector; fusing the third feature vector and the fourth feature vector to obtain a fused second feature vector; and inputting the fused feature vector into the fully-connected regression neural network (namely the regression network MLP in the figure 9), and outputting the estimated rent of the house. And meanwhile, the estimated rent, the house historical rent, a preset loss function and a preset learning rate are updated, the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network are updated, the MLP (hidden layer free) of the feature extraction network is only used for feature extraction, based on the scheme, the purpose of training the fully-connected recurrent neural network for one time is achieved, the weight in the neural network is adjusted after regression, and the accuracy of the estimated result of the fully-connected recurrent neural network is improved.
In addition, this application embodiment still provides a device is estimated to house rent, the device is estimated to house rent includes:
the acquisition module is used for acquiring house information;
and the estimation module is used for extracting a first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network and outputting the estimated rent.
In addition, the embodiment of the application also provides a terminal device, the terminal device comprises a memory, a processor and a house rent estimation program which is stored on the memory and can run on the processor, and the house rent estimation program realizes the house rent estimation method when being executed by the processor.
Since the program is executed by the processor, all technical solutions of all the foregoing embodiments are adopted, so that at least all the beneficial effects brought by all the technical solutions of all the foregoing embodiments are achieved, and details are not repeated herein.
In addition, an embodiment of the present application further provides a readable storage medium, where the readable storage medium stores a house rent estimation program, and the house rent estimation program, when executed by a processor, implements the steps of the house rent estimation method described above.
Since the program is executed by the processor, all technical solutions of all the foregoing embodiments are adopted, so that at least all the beneficial effects brought by all the technical solutions of all the foregoing embodiments are achieved, and details are not repeated herein.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrases "comprising a," "8230," "8230," or "comprising" does not exclude the presence of other like elements in a process, method, article, or system comprising the element.
The above-mentioned serial numbers of the embodiments of the present application are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present application or portions contributing to the prior art may be embodied in the form of a software product, where the computer software product is stored in a readable storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes several instructions to enable a terminal device (which may be a mobile phone, a computer, a server, a controlled terminal, or a network device) to execute the method of each embodiment of the present application.
The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are included in the scope of the present application.
Claims (10)
1. A house rent estimation method is characterized by comprising the following steps:
acquiring house information, wherein the house information comprises a house structure plan and house structural information;
and extracting a first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting a pre-estimated rent.
2. The method for predicting house rent according to claim 1, wherein the step of extracting the first feature vector from the house information and inputting the extracted first feature vector into a preset fully-connected recurrent neural network, and outputting the predicted rent further comprises:
acquiring historical house information, wherein the historical house information comprises a historical house structure plan and historical house structure information;
extracting a second feature vector from the house historical information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network;
inputting the second feature vector into the fully-connected recurrent neural network, and outputting the estimated rent of the house;
updating the weight of the fully-connected recurrent neural network, the weight of the convolutional neural network VGG16 pre-training model and the weight of the fully-connected neural network according to the estimated rent and the historical rent in the house historical information, a preset loss function and a preset learning rate;
and executing the step of acquiring the house historical information according to the preset training times, wherein the house historical information comprises a house structure historical plan and house structural historical information until the preset training times are completed.
3. The method for forecasting house rent according to claim 2, wherein the step of extracting the second feature vector from the house history information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network comprises:
extracting a third feature vector from the building structure historical plan through the convolutional neural network VGG16 pre-training model;
inputting the house structural historical information into the fully-connected neural network, and outputting a corresponding fourth feature vector;
and fusing the third feature vector and the fourth feature vector to obtain a fused second feature vector.
4. The method for estimating house rent according to claim 2, wherein the step of updating the weight of the fully-connected recurrent neural network, the weight of the pre-trained model of the convolutional neural network VGG16, and the weight of the fully-connected neural network according to the estimated rent, the historical rent in the house historical information, a preset loss function, and a preset learning rate comprises:
making a difference between the estimated rent and the rent in the house historical information, and calculating loss;
calculating a gradient according to the loss and the preset loss function to obtain a gradient descending direction;
and updating the weight of the fully-connected recurrent neural network, the weight of the pre-training model of the convolutional neural network VGG16 and the weight of the fully-connected neural network according to the gradient descent direction and the preset learning rate.
5. The method for predicting house rent according to claim 2, wherein the step of acquiring house history information according to a preset number of training times is executed, the house history information comprises the steps of house structure history plan and house structure history information, and the step of completing the preset number of training times comprises the steps of:
accumulating the training times and comparing the training times with the preset training times;
if the preset training times are reached, stopping training the fully-connected recurrent neural network;
and if the preset training times are not reached, executing the step of acquiring the house historical information until the training times reach the preset training times.
6. The method for forecasting house rent according to claim 2, wherein the step of extracting the second feature vector from the house history information through a preset convolutional neural network VGG16 pre-training model or a preset fully-connected neural network is preceded by the step of:
initializing weights of the fully-connected neural network and the fully-connected recurrent neural network;
and setting the learning rate and the training times of the fully-connected recurrent neural network.
7. The method of forecasting a house rent according to claim 2, wherein the step of obtaining house history information comprises:
and randomly acquiring the house historical information according to normal distribution.
8. A house rent estimation device, characterized in that the house rent estimation device comprises:
the acquisition module is used for acquiring house information;
and the estimation module is used for extracting the first characteristic vector from the house information, inputting the first characteristic vector into a preset full-connection regression neural network, and outputting an estimated rent.
9. A terminal device comprising a memory, a processor and a house rent estimator stored on the memory and operable on the processor, the house rent estimator being operable by the processor to perform the steps of the method of estimating a house rent according to any one of claims 1-7.
10. A readable storage medium, on which is stored a house rental estimation program, which when executed by a processor, carries out the steps of a house rental estimation method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211269958.1A CN115345684A (en) | 2022-10-18 | 2022-10-18 | House rent estimation method and device, terminal equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211269958.1A CN115345684A (en) | 2022-10-18 | 2022-10-18 | House rent estimation method and device, terminal equipment and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115345684A true CN115345684A (en) | 2022-11-15 |
Family
ID=83956998
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211269958.1A Pending CN115345684A (en) | 2022-10-18 | 2022-10-18 | House rent estimation method and device, terminal equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115345684A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109389247A (en) * | 2018-09-27 | 2019-02-26 | 智庭（北京）智能科技有限公司 | A kind of region house rent prediction technique based on big data |
| CN110084293A (en) * | 2019-04-18 | 2019-08-02 | 贝壳技术有限公司 | A kind of determination method and apparatus in complete bright pattern house |
| CN110619543A (en) * | 2019-08-23 | 2019-12-27 | 深圳市新系区块链技术有限公司 | House lease price prediction method and related device |
| CN112232885A (en) * | 2020-10-29 | 2021-01-15 | 北京工商大学 | Multi-mode information fusion-based warehouse rental price prediction method |
| CN113034242A (en) * | 2021-04-23 | 2021-06-25 | 中国建设银行股份有限公司 | Rental assistance method, device, equipment and storage medium |
| CN113204719A (en) * | 2021-04-30 | 2021-08-03 | 武汉大学 | Urban house rent assessment method based on position information superposition and deep neural network |
| KR20210158674A (en) * | 2020-06-24 | 2021-12-31 | 주식회사 더굿이너프 | System of judging outlier of real estate lease case for estimating real estate lease price |
-
2022
- 2022-10-18 CN CN202211269958.1A patent/CN115345684A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109389247A (en) * | 2018-09-27 | 2019-02-26 | 智庭（北京）智能科技有限公司 | A kind of region house rent prediction technique based on big data |
| CN110084293A (en) * | 2019-04-18 | 2019-08-02 | 贝壳技术有限公司 | A kind of determination method and apparatus in complete bright pattern house |
| CN110619543A (en) * | 2019-08-23 | 2019-12-27 | 深圳市新系区块链技术有限公司 | House lease price prediction method and related device |
| KR20210158674A (en) * | 2020-06-24 | 2021-12-31 | 주식회사 더굿이너프 | System of judging outlier of real estate lease case for estimating real estate lease price |
| CN112232885A (en) * | 2020-10-29 | 2021-01-15 | 北京工商大学 | Multi-mode information fusion-based warehouse rental price prediction method |
| CN113034242A (en) * | 2021-04-23 | 2021-06-25 | 中国建设银行股份有限公司 | Rental assistance method, device, equipment and storage medium |
| CN113204719A (en) * | 2021-04-30 | 2021-08-03 | 武汉大学 | Urban house rent assessment method based on position information superposition and deep neural network |
Non-Patent Citations (1)
| Title |
|---|
| 王军武等: "BP神经网络在城市房屋拆迁估价中的应用", 《华中科技大学学报(城市科学版)》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6386107B2 (en) | Localized learning from global models | |
| CN109145828B (en) | Method and apparatus for generating video category detection model | |
| CN109447156B (en) | Method and apparatus for generating a model | |
| US20190026630A1 (en) | Information processing apparatus and information processing method | |
| CN112269650A (en) | Task scheduling method and device, electronic equipment and storage medium | |
| KR102449630B1 (en) | Electronic device and Method for controlling the electronic device thereof | |
| CN111611085B (en) | Yun Bian collaboration-based man-machine hybrid enhanced intelligent system, method and device | |
| WO2021179818A1 (en) | Travel state recognition method and apparatus, and terminal and storage medium | |
| CN110942086A (en) | Data prediction optimization method, device and equipment and readable storage medium | |
| EP3671441A1 (en) | Application management method and apparatus, storage medium, and electronic device | |
| CN113256339B (en) | Resource release method and device, storage medium and electronic equipment | |
| CN113766633A (en) | Data processing method, data processing device, electronic equipment and storage medium | |
| CN107728772B (en) | Application processing method and device, storage medium and electronic equipment | |
| CN111158918B (en) | Supporting point parallel enumeration load balancing method, device, equipment and medium | |
| CN115345684A (en) | House rent estimation method and device, terminal equipment and readable storage medium | |
| CN110610140B (en) | Training method, device and equipment of face recognition model and readable storage medium | |
| CN109720945B (en) | Elevator allocation method, device, equipment and computer readable storage medium | |
| CN113034580B (en) | Image information detection method and device and electronic equipment | |
| CN109436980B (en) | Method and system for detecting state of elevator component | |
| CN112948763B (en) | Piece quantity prediction method and device, electronic equipment and storage medium | |
| CN115016950A (en) | Data analysis method and system based on multithreading model | |
| CN114446427A (en) | Electronic equipment and health data attribution identification method | |
| CN112328769A (en) | Automatic customer service response method, device and computer readable storage medium | |
| CN112529602A (en) | Data processing method and device, readable storage medium and electronic equipment | |
| CN114723469A (en) | Method, system and electronic device for user satisfaction degree prediction and attribution |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221115 |
|
| RJ01 | Rejection of invention patent application after publication |