WO2022198963A1 - Big data-based commercial space quality evaluation method and system, device, and medium - Google Patents

Abstract

A big data-based commercial space quality evaluation method and system, device, and medium. The method comprises: obtaining POI data and DianPing data within a research range, generating a plurality of street view acquisition points and obtaining corresponding street view image data; calculating two indexes to represent objective evaluation; obtaining the compelling degree indicator of commercial facilities according to the DianPing data; building a random forest model, and using a trained random forest model to score second-level indicators of the street view image data of individual street view acquisition points to obtain scores of three indicators, which constitute first-level indicators together with the compelling degree of the commercial facilities to represent subjective evaluation; and performing calculation on the basis of the objective evaluation indicators and the scores of the subjective evaluation indicators of individual street view acquisition points to achieve comprehensive evaluation of the commercial space quality. The method integrates objective and subjective evaluations of commercial space quality and employs the machine learning algorithm to achieve large-scale and high-precision assessment of commercial space quality.

Description

Commercial space quality evaluation method, system, equipment and medium based on big data technical field

The invention relates to the field of commercial space quality evaluation, in particular to a big data-based commercial space quality evaluation method, system, computer equipment and storage medium.

Background technique

Commercial space is a highly dynamic area of the city, attracting large crowds. As cities grow, traditional commercial spaces face various problems, such as traffic congestion, aging buildings, and declining vitality. In recent years, the rapid development of mobile information and communication technology and the popularization of smart phones have accelerated the development and penetration of online shopping, which has brought a huge impact on urban commercial space. The improvement of the quality of urban commercial space has become the key to maintaining its competitiveness.

Most of the existing researches on the quality of commercial space use traditional questionnaires, field surveys, etc. These methods are suitable for evaluating the quality of small-scale commercial space, but it is difficult to achieve accurate commercial space quality in a larger area or even the entire city. measure. On the other hand, the measurement of commercial space quality should not only include objective commercial facilities, but also the needs and feelings of subjective individuals. Most of the existing research focuses on one aspect, and rarely combines subjective and objective elements to comprehensively measure the city. Commercial space quality.

SUMMARY OF THE INVENTION

In order to solve the above-mentioned deficiencies of the prior art, the present invention provides a big data-based commercial space quality evaluation method, system, computer equipment and storage medium, which can be used to evaluate a wide range of commercial space quality, Combining elements, comprehensively evaluate the quality of commercial space, so that the evaluation results are more in line with objective reality.

The first object of the present invention is to provide a commercial space quality evaluation method based on big data.

The second object of the present invention is to provide a commercial space quality evaluation system based on big data.

A third object of the present invention is to provide a computer device.

A fourth object of the present invention is to provide a storage medium.

The first purpose of the present invention can be achieved by adopting the following technical solutions:

A method for evaluating the quality of commercial space based on big data, the method comprising:

Obtain POI data and public comment data within the scope of the study;

According to the POI data, generate several street view collection points, and obtain corresponding street view image data at each street view collection point;

According to the POI data, calculate the traffic convenience index and the commercial format richness index of each street view collection point to represent the objective evaluation of the commercial space quality;

According to the public comment data, the total number of evaluations of each street view collection point is counted, and the attention index of commercial facilities is obtained;

constructing a random forest model, and using the data generated from the street view image data to train the random forest model;

Using the trained random forest model, evaluate the score values of the secondary indicators of the street view image data of each street view collection point, and obtain the score values of the commercial street space, commercial space atmosphere, and commercial building facade indicators, respectively, Together with the degree of attention of the commercial facilities, it constitutes a first-level indicator, which represents the subjective evaluation of the quality of commercial space;

According to the objective evaluation of commercial space quality and the subjective evaluation of commercial space quality, the comprehensive evaluation of commercial space quality is calculated.

Further, generating several street view collection points according to the POI data, and obtaining corresponding street view image data at each street view collection point, specifically including:

According to POI data, determine the business area to be evaluated within the research scope;

In the commercial area, several street view collection points are generated, and corresponding street view image data is acquired at each street view collection point.

Further, according to the POI data, the business area to be evaluated within the research scope is determined, specifically including:

Select commercial facility POIs from the POI data, perform cluster analysis on the commercial facility POIs, and classify commercial facility POIs into several categories according to the distribution density of commercial facility POIs;

The POI points of the same cluster are combined into one area as the commercial area to be evaluated.

Further, generating several street view collection points in the commercial area, and acquiring corresponding street view image data at each street view collection point, specifically including:

Select road network data in each commercial area;

based on the road network data, generating several street view collection points at set intervals;

According to the direction of the road where each street view collection point is located, street view images from four horizontal perspectives parallel and perpendicular to the road direction are obtained respectively, and no street view images and duplicate collection points are eliminated.

Further, according to the POI data, the traffic convenience index and the commercial format richness index of each street view collection point are calculated, specifically including:

According to POI data, obtain the spatial distribution of bus stations and subway stations and the spatial distribution of various commercial facilities;

Summarize the number of bus stops and subway stations of each street view collection point, and calculate the traffic convenience index of each street view collection point:

Among them, Q _i is the number of bus stops of the i-th street view collection point within the first preset range, Q _max is the maximum number of bus stops, and Ri is the number of subway stations of the i-th street view collection point within the second preset range , Rmax is the maximum number of subway stations; a, b represent the importance weights of bus stations and subway stations, respectively;

Count the number of commercial formats of each street view collection point within the third preset range, and use Shannon entropy to calculate the commercial format mixture of each street view collection point:

Among them, m is the number of commercial formats, which include catering, shopping, finance, life services, leisure and entertainment; T _j is the number of commercial facilities POIs of commercial format type j accounting for the ith street view collection point in the third preset range Proportion of the number of POIs in all commercial facilities;

Calculate the commercial richness index of each Street View collection point:

Among them, P _i is the commercial format mixture degree of the ith street view collection point, P _max is the maximum commercial format mixture degree; U _i is the number of commercial facilities POIs of the ith street view collection point within the third preset range, U _max is the maximum number of commercial facility POIs; x and y represent the weight of the commercial format mix and the number of commercial facilities, respectively.

Further, the random forest model is constructed, and the random forest model is trained by using the data generated from the street view image data, which specifically includes:

Semantic segmentation is performed on the street view image data by using a fully convolutional neural network to obtain feature data of the street view image data;

Randomly select part of the street view image data, and obtain the scores entered by volunteers for the secondary indicators of the street view image data as the scoring sample data; among them, the secondary indicators of the street view image data include the cleanliness of commercial streets and the suitability of the aspect ratio of commercial streets. , commercial street openness, commercial street walkability, commercial street greening degree, commercial space atmosphere, commercial building facade cleanliness;

Build a random forest model, and randomly select part of the data and the corresponding street view image feature data from the scoring sample data as the training set;

In the training set, the random forest model is trained with the scoring sample data as the dependent variable and the street view image feature data as the independent variable;

When the model performance metrics reach the standard value, the training of the random forest model is completed.

Further, the trained random forest model is used to evaluate the score value of the secondary indicators of the street view image data of each street view collection point, and obtain the indicators of commercial street space, commercial space atmosphere, and commercial building facade respectively. The score value of , specifically:

Use the trained random forest model to evaluate the score value of the secondary indicators of the street view image data of each street view collection point. The score value of the indicator.

Further, according to the objective evaluation of the commercial space quality and the subjective evaluation of the commercial space quality, the comprehensive evaluation of the commercial space quality is calculated and obtained, specifically:

According to statistics, the evaluation weights of the traffic convenience index K and the business richness index D are w _Ki , w _Di , and the evaluation weights of the commercial street space STR, the commercial space atmosphere ENV, the commercial building facade BUI, and the commercial facility attention ATT are s _1i , s _2i , s _3i , s _4i , the importance weights of the objective evaluation index and the subjective evaluation index are g _1i , g _2i ;

Dimensionless processing is adopted for the commercial street space index, commercial space atmosphere index, commercial building facade index and commercial facility attention index, as follows:

Among them, STR _i , ENV _i , BUI _i , and ATT _i are the score values of the commercial street space index, commercial space atmosphere index, commercial building facade index, and commercial facility attention index of the i-th street view collection point, respectively, STR _max , ENV _max , BUI _max , ATT _max respectively represent the maximum value of the corresponding index score;

According to the following formula, the comprehensive evaluation value of the commercial space quality of the street view collection point i is calculated:

C _i =g _1i (w _Ki K _i +w _Di D _i )+g _2i (s _1i O _STRi +s _2i O _ENVi +s _3i O _BUIi +s _4i O _ATTi )×100

Among them, K _i and D _i are the traffic convenience index and the business richness index of the i-th street view collection point, respectively.

The second object of the present invention can be achieved by adopting the following technical solutions:

A commercial space quality evaluation system based on big data, the system includes:

The data acquisition module is used to acquire POI data and public comment data within the research scope;

Generate a street view collection point module for generating several street view collection points according to the POI data, and obtain corresponding street view image data at each street view collection point;

The objective evaluation module is used to calculate the traffic convenience index and the commercial format richness index of each street view collection point according to the POI data, so as to represent the objective evaluation of the commercial space quality;

The attention degree module of commercial facilities is used to count the sum of the evaluation quantity of each street view collection point according to the public comment data, and obtain the attention degree index of commercial facilities;

constructing a random forest module for constructing a random forest model, and using the data generated from the street view image data to train the random forest model;

The subjective evaluation module is used to use the trained random forest model to evaluate the score value of the secondary indicators of the street view image data of each street view collection point, and obtain the commercial street space, commercial space atmosphere, and commercial building facade respectively. The score value of the indicator, together with the degree of attention of the commercial facilities, constitute the first-level indicator, which represents the subjective evaluation of the quality of commercial space;

The comprehensive evaluation module is used to calculate the comprehensive evaluation of the commercial space quality according to the objective evaluation of the commercial space quality and the subjective evaluation of the commercial space quality.

The third object of the present invention can be achieved by adopting the following technical solutions:

A computer device includes a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the above-mentioned commercial space quality evaluation method is implemented.

The fourth object of the present invention can be achieved by adopting the following technical solutions:

A storage medium storing a program, when the program is executed by a processor, realizes the above-mentioned commercial space quality evaluation method.

The present invention has the following beneficial effects with respect to the prior art:

The invention adopts the combination of machine learning algorithm and street view image to construct a random forest model for commercial space quality assessment, which is conducive to realizing a large-scale and high-precision measurement of urban commercial space quality; It evaluates the quality of commercial space at the level, and integrates the subjective perception of micro-individuals on the quality of commercial space, that is, comprehensively evaluates the quality of commercial space from two dimensions, subjective and objective, so that the evaluation results are more in line with objective reality.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can also be obtained according to the structures shown in these drawings without creative efforts.

FIG. 1 is a flow chart of the method for evaluating the quality of commercial space based on big data according to Embodiment 1 of the present invention.

FIG. 2 is a distribution diagram of a commercial area in Yuexiu District according to Embodiment 1 of the present invention.

FIG. 3 is a spatial distribution diagram of the traffic convenience index in Yuexiu District according to Embodiment 1 of the present invention.

FIG. 4 is a spatial distribution diagram of the commercial richness index of Yuexiu District in Example 1 of the present invention.

FIG. 5 is a spatial distribution diagram of various subjective evaluation indicators in Yuexiu District according to Embodiment 1 of the present invention.

6 is a spatial distribution diagram of comprehensive evaluation of commercial space quality in Yuexiu District according to Embodiment 1 of the present invention.

FIG. 7 is a structural block diagram of a commercial space quality evaluation system according to Embodiment 2 of the present invention.

FIG. 8 is a structural block diagram of a computer device according to Embodiment 3 of the present invention.

Detailed ways

In order to make the purposes, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present invention. .

Example 1:

This embodiment takes Yuexiu District, Guangzhou City as an example, and provides a big data-based commercial space quality evaluation method, which uses a combination of multi-source big data and machine learning algorithms to construct a model for evaluating commercial space quality, and accurately evaluates large-scale commercial space quality. Urban commercial space quality.

As shown in Figure 1, the big data-based commercial space quality evaluation method includes the following steps:

S1, data acquisition and processing.

Firstly, the POI data of Yuexiu District, Guangzhou City, OSM road network data and public comment data were obtained, and the POI of commercial facilities was screened from the POI data, and the commercial area to be evaluated within the research scope was further analyzed and determined. The road network data in the commercial area is selected, and the locations of the street view collection points are determined at specific intervals along the road network. Based on the location coordinates of the collection point, the street view image data is crawled from the Baidu map open platform.

Further, the specific implementation of step S1 is as follows:

Screen commercial facility POIs from the acquired POI data, and use ArcGIS Pro software to perform DBSCAN cluster analysis on the commercial facility POIs. The POI points of the class are combined into a region. Specifically, since the POI distribution density of city-level commercial areas is much higher than that of community-level commercial areas, in order to avoid the loss of community-level commercial areas, two clustering analyses were performed successively. First, perform DBSCAN cluster analysis on POIs of all commercial facilities, set the minimum number of cluster elements as e, and obtain the POI clustering of the core business area; In the secondary DBSCAN cluster analysis, the minimum number of cluster elements is set to f (e>f), and the POI clustering of the sub-core commercial area is obtained. On this basis, the Euclidean allocation tool is used to convert the two clustering results into area, which is used as the commercial area of this study.

Specifically, in Yuexiu District, Guangzhou, the distribution of its commercial areas is shown in Figure 2. The core commercial areas are mainly distributed in Beijing Road business district, Huanshidong business district, China Plaza, Sanyuanli and Wuyang Village, and the second The core commercial areas are mostly distributed near the core commercial areas, such as Dongxianli Community, Panfu Community on Liurong Street, and the community on the west side of Xiaobei Road.

The street view image in this embodiment comes from the Baidu map open platform (http://lbsyun.baidu.com), and the road network data in each commercial area is selected according to the scope of the commercial area. Based on the road network data, using the densification tool in ArcGIS Pro, several street view collection points are generated at specific intervals. According to the direction of the road where each collection point is located, the street view images of four horizontal perspectives parallel and perpendicular to the direction of the road are obtained respectively. After eliminating no street view images and duplicate collection points, 9424 street view images from 2356 collection points in the commercial area of Yuexiu District, Guangzhou City were finally obtained.

S2, objective evaluation.

Measure the traffic convenience index and commercial format richness index in the commercial area, that is, use the POI data of Yuexiu District, Guangzhou City to obtain the spatial distribution of public transportation stations (including bus stations and subway stations) and the space of various commercial facilities. The degree of mixing represents an objective evaluation of the quality of commercial spaces.

Further, the specific implementation of step S2 is as follows:

In the objective evaluation of commercial space quality, starting from the transportation convenience index and commercial richness index, firstly, the number of bus stops within the 500-meter buffer area of each collection point and the subway stations within the 1,000-meter buffer area of each collection point are summarized respectively. number, and then calculate the transportation convenience index of each collection point:

Among them, Qi is the number of bus stops within the 500m buffer of the _ith collection point; _Qmax is the maximum number of bus stops; Ri is the number of subway stations within the 1000m buffer of the ith collection point; Rmax is the maximum subway station Points; where a and b represent the importance weights of bus stops and subway stations in commercial districts, respectively, using a network questionnaire. Statistics show that bus: subway = a: b = 0.08: 0.92. On this basis, the spatial distribution of the transportation convenience index is represented by kernel density analysis.

As shown in Figure 3, the areas with the best transportation convenience are mainly distributed in the Memorial Hall, the north side of Tuanyi Square, Dashatou Sanma Road and Guzhuang, followed by Post and Telecommunications Community, the north side of the intersection of Jiefang South Road and Daxin Road , agricultural lecture center and other areas. It is worth noting that most Guangzhou residents believe that the importance of the subway stations around the business district is much higher than that of the bus station, so the commercial areas near the subway station are generally more convenient for transportation.

Further, count the number of commercial format types within the 100-meter buffer area of each collection point, and use Shannon entropy to calculate the commercial format mixing degree of each collection point:

Among them, m is the number of types of commercial formats, including the types of commercial formats j are catering, shopping, finance, life services, leisure and entertainment, etc.; T _j is the number of POIs of facilities of commercial type j accounting for the 100-meter buffer area of the i-th collection point Proportion of the number of POIs in all commercial facilities. On this basis, calculate the commercial richness index of each collection point:

Among them, Pi is the commercial format mixing degree of the _ith collection point, _Pmax is the maximum commercial format mixing degree; U _i is the number of commercial facilities POI within the 100-meter buffer area of the ith collection point, and _Umax is the maximum commercial facility. The number of POIs; where x and y represent the weights of the commercial format mixture degree and the number of commercial facilities obtained by the online questionnaire respectively; the statistics show that the commercial format mixture degree: the number of commercial facilities = x: y = 0.89: 0.11. On this basis, the spatial distribution of commercial richness indices was represented by kernel density analysis.

As shown in Figure 4, the high-value areas of the business richness index are mainly distributed in Memorial Hall, Lujing Road, Wuyang Village, the intersection of Huanshi East Road and Xianlie South Road, and near Dashatou Sanma Road, while Beijing Road Merchants The business richness index of the districts such as China Plaza, China Plaza and Sanyuanli is relatively low.

S3, machine learning.

The fully convolutional neural network is used to semantically segment the acquired street view image data to obtain the feature data of the street view image. Part of the street view image data is randomly selected for manual scoring, and the scoring sample data is obtained.

Build a random forest model for commercial space quality assessment, take 75% of the scoring sample data and the corresponding street view image feature data as the training set, and the remaining 25% of the scoring sample data and street view image feature data as the prediction set; among them, the scoring sample data as The dependent variable, the image feature data is used as the independent variable, and the random forest model is trained.

Further, the specific implementation of step S3 is as follows:

First, the fully convolutional neural network FCN-8s (Yao Y, et al., 2019) developed by the CUG.HPSCIL laboratory is used to semantically segment the acquired street view images. It is predicted to be one of 151 types of ground objects including "unknown", and then the segmented PNG image file and image feature statistics csv file (including image feature data) are obtained. Second, four volunteers with a professional background in architecture or urban and rural planning were selected. Each volunteer randomly selected 800 street view images from 200 collection points, and gave a score for each indicator according to their own perception of the quality of commercial space. The indicators for evaluating the quality of commercial spaces determined in this embodiment include the cleanliness of commercial streets, the suitability of the height-to-width ratio of commercial streets, the openness of commercial streets, the walkability of commercial streets, the degree of greening of commercial streets, the atmosphere of commercial spaces, and the standing of commercial buildings. There are 87 secondary indicators of surface cleanliness, and the scoring range of each indicator is set to [0, n], where 0 is the lowest score, n is the highest score, and n is greater than 0. Finally, the scoring results of 800 collection points are obtained as the scoring sample data.

Further, build a random forest model for commercial space quality evaluation, and use the obtained 800 scoring sample data as the data set for training and prediction models;

Specifically, 75% of the commercial space quality index scoring results of street view images and their corresponding street view image feature data are randomly selected from the scoring sample data as the training set, and the remaining 25% of the street view image scoring results and street view image feature data are used as test set. Further, the commercial space quality index score is used as the dependent variable, and the street view image feature data is used as the independent variable to train the random forest model for commercial space quality evaluation. When the model performance metrics reach the standard value (specifically, the classification accuracy reaches 0.8 and the Kappa coefficient ≥ 0.6), it means that the training of the random forest model is completed.

Wherein, the model performance metrics are the classification accuracy and Kappa coefficient of the commercial space quality assessment random forest model on the test set, and the classification accuracy refers to the proportion of correctly scored street view images to all images in the test set, and the classification accuracy Reaching 0.8 indicates that the accuracy of the prediction results is high; the Kappa coefficient is used to evaluate the degree of correlation between the prediction results and the volunteer scoring results, and the Kappa coefficient ≥ 0.6 indicates that the correlation between the prediction results and the scoring results is good.

S4, subjective evaluation.

According to the evaluation data of Dianping, the attention index of commercial facilities is measured.

Use the random forest model trained in step S3 to evaluate the scores of 87 secondary indicators of street view images in the commercial area, and further sum up to obtain the score values of commercial street space, commercial space atmosphere, and commercial building facade indicators, which are closely related to commercial facilities. The degree indicators together constitute four first-level indicators, which represent the individual's subjective evaluation of the quality of commercial space.

Further, the specific implementation of step S4 is as follows:

Based on the obtained Dianping evaluation data, the total number of evaluations of Dianping stores within the 100-meter buffer area of each collection point is counted to represent the attention of commercial facilities.

Use the trained random forest model to evaluate the scores of 7 secondary indicators in the commercial area on a large scale, and further add the scores of commercial street space, commercial space atmosphere, commercial building facade indicators, and commercial facilities attention. Together, they constitute four primary indicators, which represent the individual's subjective evaluation of the quality of commercial space. On this basis, the spatial distribution of the first-level index scores of each subjective evaluation is represented by kernel density analysis.

As shown in Figure 5, the area with the highest attention on commercial facilities is Beijing Road business district, followed by Huanshi East business district and the north side of Tuanyi Square, both of which have the same level of attractiveness. The commercial street space, commercial space atmosphere and commercial space in Guangzhou Railway Station, Lujing Road, Huanshi East Business Circle, Memorial Hall, Xiaobei Road, Dashatou Sanma Road, Wuyang Village and the north side of Tuanyi Square are The scores of building facade indicators are all high. Based on the scores of the above indicators, the areas with higher overall subjective evaluation of commercial space quality are Guangzhou Railway Station, Lujing Road, Huanshi East Business Circle, Memorial Hall, and Dashatou Sanma Road.

S5, comprehensive evaluation.

The subjective and objective commercial space quality indicators of commercial facilities are weighted and calculated to form a comprehensive evaluation chart for commercial space quality evaluation. In terms of specific calculation, the two objective evaluation indicators and the first-level indicators of the four subjective evaluations described in step S4 are adopted. The weight is determined by the online questionnaire; the comprehensive weight is further determined in the same way.

Further, the specific implementation of step S5 is as follows:

First, according to the statistics of the online questionnaire survey, the evaluation weights w _Ki and w _Di of the traffic convenience index (K) and the business richness index (D) are obtained, the commercial street space (STR), the commercial space atmosphere (ENV), the commercial building The evaluation weights s _1i , s _2i , s _3i , and s _4i of face (BUI) and commercial facility attention (ATT). According to statistics, w _Ki : w _Di = 0.47: 0.53; s _1i : s _2i : s _3i : s _4i = 0.36: 0.28: 0.20: 0.16. The importance weight values g _1i and g _2i of the objective evaluation index and the subjective evaluation index are obtained by means of network questionnaire survey, and g _1i : g _2i =0.56:0.44 is obtained through statistics.

Secondly, due to the large difference in the order of magnitude of each subjective first-level index, dimensionless processing is adopted for each subjective first-level index:

Among them, STR _i , ENV _i , BUI _i , and ATT _i are the scores of the ith collection point commercial street space, commercial space atmosphere, commercial building facade, and commercial facilities attention indicators, respectively, STR _max , ENV _max , BUI _max , ATT _max represents the maximum value of the corresponding index score, respectively.

From this, the comprehensive evaluation value of the commercial space quality of the collection point i is obtained:

C _i =g _1i (w _Ki K _i +w _Di D _i )+g _2i (s _1i O _STRi +s _2i O _ENVi +s _3i O _BUIi +s _4i O _ATTi )×100

Among them, K _i and D _i are the traffic convenience index and the business richness index of the i-th street view collection point, respectively.

According to this method, a subjective and objective comprehensive evaluation is carried out on the evaluation of the commercial space quality of Yuexiu District. The comprehensive evaluation results are shown in Figure 6. The high-value areas of the comprehensive evaluation are concentrated in the Memorial Hall, Lujing Road, Quzhuang, and Tuanyi University. The north side of the square and Dashatou Sanma Road and other areas.

To sum up, the present invention explores the application of commercial space quality evaluation based on big data by selecting typical commercial areas for data mining, street view image semantic segmentation, and machine learning. In terms of objective evaluation, the obtained multi-source big data is used to evaluate the traffic convenience index and commercial richness index of commercial areas; in terms of subjective evaluation, a seven-dimensional index system is constructed, and street view images are manually scored from these dimensions. Using machine learning algorithms, a random forest model for commercial space quality evaluation is constructed to predict the subjective evaluation of large-scale commercial space quality. On the basis of the above, weights are calculated for the subjective and objective evaluations of commercial space to form a comprehensive evaluation of commercial space quality scores. It not only evaluates the quality of commercial space from an objective level, but also integrates the subjective perception of micro-individuals on the quality of commercial space. At the same time, the use of machine learning algorithms is conducive to the realization of large-scale and high-precision measurement of urban commercial space quality.

Those skilled in the art can understand that all or part of the steps in the method for implementing the above embodiments can be completed by instructing relevant hardware through a program, and the corresponding program can be stored in a computer-readable storage medium.

It should be noted that although the method operations of the above-described embodiments are depicted in a particular order in the drawings, this does not require or imply that the operations must be performed in that particular order, or that all illustrated operations must be performed to achieve the desired results . Conversely, the depicted steps may change the order of execution. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined to be performed as one step, and/or one step may be decomposed into multiple steps to be performed.

Example 2:

As shown in FIG. 7 , this embodiment provides a big data-based commercial space quality evaluation system. The system includes a data acquisition module 701 , a street view collection point generation module 702 , an objective evaluation module 703 , and a commercial facility attention module 704 , Build a random forest module 705, a subjective evaluation module 706 and a comprehensive evaluation module 707. The specific functions of each module are as follows:

A data acquisition module 701 is used to acquire POI data and public comment data within the research scope;

generating a street view collection point module 702, configured to generate several street view collection points according to the POI data, and obtain corresponding street view image data at each street view collection point;

The objective evaluation module 703 is used to calculate the traffic convenience index and the commercial format richness index of each street view collection point according to the POI data, to represent the objective evaluation of the commercial space quality;

The attention degree module 704 of commercial facilities is configured to count the sum of the evaluation quantity of each street view collection point according to the public comment data, and obtain the attention degree index of commercial facilities;

Building a random forest module 705 for building a random forest model, and using the data generated from the street view image data to train the random forest model;

The subjective evaluation module 706 is configured to use the trained random forest model to evaluate the score values of the secondary indicators of the street view image data of each street view collection point, and obtain respectively the commercial street space, commercial space atmosphere, and commercial building standing. The score value of the surface index, together with the attention of the commercial facilities, constitute the first-level index, which represents the subjective evaluation of the quality of commercial space;

The comprehensive evaluation module 707 is configured to obtain a comprehensive evaluation of the commercial space quality according to the objective evaluation of the commercial space quality and the subjective evaluation of the commercial space quality, and calculate the comprehensive evaluation of the commercial space quality.

For the specific implementation of each module in this embodiment, reference may be made to the above-mentioned Embodiment 1, which will not be repeated here. It should be noted that the system provided in this embodiment only takes the division of the above-mentioned functional modules as an example, and in practical applications , the above-mentioned function distribution can be completed by different function modules according to the needs, that is, the internal structure is divided into different function modules, so as to complete all or part of the functions described above.

Example 3:

This embodiment provides a computer device, which may be a computer, as shown in FIG. 8 , which includes a processor 802 , a memory, an input device 803 , a display 804 and a network interface 805 connected through a system bus 801 , the processing The memory is used to provide computing and control capabilities, the memory includes a non-volatile storage medium 806 and an internal memory 807 that stores an operating system, computer programs and databases, and the internal memory 807 is non-volatile The operating system and the computer program in the storage medium provide an environment, and when the processor 802 executes the computer program stored in the memory, it implements the commercial space quality evaluation method of the above-mentioned embodiment 1, as follows:

Obtain POI data and public comment data within the scope of the study;

According to the POI data, generate several street view collection points, and obtain corresponding street view image data at each street view collection point;

According to the POI data, calculate the traffic convenience index and the commercial format richness index of each street view collection point to represent the objective evaluation of the commercial space quality;

According to the public comment data, the total number of evaluations of each street view collection point is counted, and the attention index of commercial facilities is obtained;

constructing a random forest model, and using the data generated from the street view image data to train the random forest model;

Using the trained random forest model, evaluate the score values of the secondary indicators of the street view image data of each street view collection point, and obtain the score values of the commercial street space, commercial space atmosphere, and commercial building facade indicators, respectively, Together with the degree of attention of the commercial facilities, it constitutes a first-level index, which represents the subjective evaluation of the commercial space quality; according to the objective evaluation of the commercial space quality and the subjective evaluation of the commercial space quality, the comprehensive evaluation of the commercial space quality is calculated.

Example 4:

This embodiment provides a storage medium, which is a computer-readable storage medium, and stores a program. When the program is executed by a processor, the method for evaluating the commercial space quality of the above-mentioned Embodiment 1 is implemented, as follows:

Obtain POI data and public comment data within the scope of the study;

According to the POI data, generate several street view collection points, and obtain corresponding street view image data at each street view collection point;

According to the POI data, calculate the traffic convenience index and the commercial format richness index of each street view collection point to represent the objective evaluation of the commercial space quality;

According to the public comment data, the total number of evaluations of each street view collection point is counted, and the attention index of commercial facilities is obtained;

constructing a random forest model, and using the data generated from the street view image data to train the random forest model;

Using the trained random forest model, evaluate the score values of the secondary indicators of the street view image data of each street view collection point, and obtain the score values of the commercial street space, commercial space atmosphere, and commercial building facade indicators, respectively, Together with the degree of attention of the commercial facilities, it constitutes a first-level indicator, which represents the subjective evaluation of the quality of commercial space;

According to the objective evaluation of commercial space quality and the subjective evaluation of commercial space quality, the comprehensive evaluation of commercial space quality is calculated.

The storage medium in the above-mentioned embodiment may be a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a U disk, a removable hard disk and other media.

To sum up, the present invention uses a combination of machine learning algorithms and street view images to construct a random forest model for commercial space quality assessment, which is conducive to the realization of a large-scale and high-precision measurement of urban commercial space quality; The data not only evaluates the quality of commercial space from an objective level, but also integrates the subjective perception of micro-individuals on the quality of commercial space, that is, comprehensively evaluates the quality of commercial space from both subjective and objective dimensions, so that the evaluation results are more in line with objective reality.

The above is only a preferred embodiment of the patent of the present invention, but the protection scope of the patent of the present invention is not limited to this. The technical solution and the inventive concept of the invention are equivalently replaced or changed, all belong to the protection scope of the patent of the present invention.

Claims (10)

1. A method for evaluating the quality of commercial space based on big data, characterized in that the method includes:

   Obtain POI data and public comment data within the scope of the study;

   According to the POI data, generate several street view collection points, and obtain corresponding street view image data at each street view collection point;

   According to the POI data, calculate the traffic convenience index and the commercial format richness index of each street view collection point to represent the objective evaluation of the commercial space quality;

   According to the public comment data, the total number of evaluations of each street view collection point is counted, and the attention index of commercial facilities is obtained;

   constructing a random forest model, and using the data generated from the street view image data to train the random forest model;

   Using the trained random forest model, evaluate the score values of the secondary indicators of the street view image data of each street view collection point, and obtain the score values of the commercial street space, commercial space atmosphere, and commercial building facade indicators, respectively, Together with the degree of attention of the commercial facilities, it constitutes a first-level indicator, which represents the subjective evaluation of the quality of commercial space;

   According to the objective evaluation of commercial space quality and the subjective evaluation of commercial space quality, the comprehensive evaluation of commercial space quality is calculated.
2. The commercial space quality evaluation method according to claim 1, wherein generating several street view collection points according to POI data, and acquiring corresponding street view image data at each street view collection point, specifically includes:

   According to POI data, determine the business area to be evaluated within the research scope;

   In the commercial area, several street view collection points are generated, and corresponding street view image data is acquired at each street view collection point.
3. The commercial space quality evaluation method according to claim 2, characterized in that, determining the commercial area to be evaluated within the research scope according to POI data, specifically comprising:

   Select commercial facility POIs from the POI data, perform cluster analysis on the commercial facility POIs, and classify commercial facility POIs into several categories according to the distribution density of commercial facility POIs;

   The POI points of the same cluster are combined into one area as the commercial area to be evaluated.
4. The commercial space quality evaluation method according to claim 2, wherein generating several street view collection points in the commercial area, and acquiring corresponding street view image data at each street view collection point, specifically includes:

   Select road network data in each commercial area;

   based on the road network data, generating several street view collection points at set intervals;

   According to the direction of the road where each street view collection point is located, street view images from four horizontal perspectives parallel and perpendicular to the road direction are obtained respectively, and no street view images and duplicate collection points are eliminated.
5. The commercial space quality evaluation method according to claim 1, wherein the calculation of the traffic convenience index and the commercial format richness index of each street view collection point according to POI data specifically includes:

   According to POI data, obtain the spatial distribution of bus stations and subway stations and the spatial distribution of various commercial facilities;

   Summarize the number of bus stops and subway stations of each street view collection point, and calculate the traffic convenience index of each street view collection point:

   

   Among them, Q _i is the number of bus stops of the i-th street view collection point within the first preset range, Q _max is the maximum number of bus stops, and Ri is the number of subway stations of the i-th street view collection point within the second preset range , Rmax is the maximum number of subway stations; a, b represent the importance weights of bus stations and subway stations, respectively;

   Count the number of commercial formats of each street view collection point within the third preset range, and use Shannon entropy to calculate the commercial format mixture of each street view collection point:

   

   Among them, m is the number of commercial formats, which include catering, shopping, finance, life services, leisure and entertainment; T _j is the number of commercial facilities POIs of commercial format type j accounting for the ith street view collection point in the third preset range Proportion of the number of POIs in all commercial facilities;

   Calculate the commercial richness index of each Street View collection point:

   

   Among them, P _i is the commercial format mixture degree of the ith street view collection point, P _max is the maximum commercial format mixture degree; U _i is the number of commercial facilities POIs of the ith street view collection point within the third preset range, U _max is the maximum number of commercial facility POIs; x and y represent the weight of the commercial format mix and the number of commercial facilities, respectively.
6. The commercial space quality evaluation method according to claim 1, wherein the building a random forest model, and using the data generated from the street view image data to train the random forest model, specifically includes:

   Using a fully convolutional neural network to perform semantic segmentation on the street view image data to obtain feature data of the street view image data;

   Randomly select part of the street view image data, and obtain the scores entered by volunteers for the secondary indicators of the street view image data as the scoring sample data; among them, the secondary indicators of the street view image data include the cleanliness of commercial streets and the suitability of the aspect ratio of commercial streets. , commercial street openness, commercial street walkability, commercial street greening degree, commercial space atmosphere, commercial building facade cleanliness;

   Build a random forest model, and randomly select part of the data and the corresponding street view image feature data from the scoring sample data as the training set;

   In the training set, the random forest model is trained with the scoring sample data as the dependent variable and the street view image feature data as the independent variable;

   When the model performance metrics reach the standard value, the training of the random forest model is completed.
7. The commercial space quality evaluation method according to claim 1, wherein the trained random forest model is used to evaluate the score values of the secondary indicators of the street view image data of each street view collection point, respectively Obtain the score values of commercial street space, commercial space atmosphere, and commercial building facade indicators, specifically:

   Use the trained random forest model to evaluate the score value of the secondary indicators of the street view image data of each street view collection point. The score value of the indicator.
8. The commercial space quality evaluation method according to claim 1, wherein the comprehensive evaluation of the commercial space quality is calculated according to the objective evaluation of the commercial space quality and the subjective evaluation of the commercial space quality, specifically:

   According to statistics, the evaluation weights of the traffic convenience index K and the business richness index D are w _Ki , w _Di , and the evaluation weights of the commercial street space STR, the commercial space atmosphere ENV, the commercial building facade BUI, and the commercial facility attention ATT are s _1i , s _2i , s _3i , s _4i , the importance weights of the objective evaluation index and the subjective evaluation index are g _1i , g _2i ;

   Dimensionless processing is adopted for the commercial street space index, commercial space atmosphere index, commercial building facade index and commercial facility attention index, as follows:

   

   

   

   

   Among them, STR _i , ENV _i , BUI _i , and ATT _i are the score values of the commercial street space index, commercial space atmosphere index, commercial building facade index, and commercial facility attention index of the i-th street view collection point, respectively, STR _max , ENV _max , BUI _max , ATT _max respectively represent the maximum value of the corresponding index score;

   According to the following formula, the comprehensive evaluation value of the commercial space quality of the street view collection point i is calculated:

   C _i =g _1i (w _Ki K _i +w _Di D _i )+g _2i (s _1i O _STRi +s _2i O _ENVi +s _3i O _BUIi +s _4i O _ATTi )×100

   Among them, K _i and D _i are the traffic convenience index and the business richness index of the i-th street view collection point, respectively.
9. A commercial space quality evaluation system based on big data, characterized in that the system includes:

   The data acquisition module is used to acquire POI data and public comment data within the research scope;

   generating a street view collection point module for generating several street view collection points according to the POI data, and obtaining corresponding street view image data at each street view collection point;

   The objective evaluation module is used to calculate the traffic convenience index and the commercial format richness index of each street view collection point according to the POI data, so as to represent the objective evaluation of the commercial space quality;

   The attention degree module of commercial facilities is used to count the sum of the evaluation quantity of each street view collection point according to the public comment data, and obtain the attention degree index of commercial facilities;

   constructing a random forest module for constructing a random forest model, and using the data generated from the street view image data to train the random forest model;

   The subjective evaluation module is used to use the trained random forest model to evaluate the score value of the secondary indicators of the street view image data of each street view collection point, and obtain the commercial street space, commercial space atmosphere, and commercial building facade respectively. The score value of the indicator, together with the degree of attention of the commercial facilities, constitute the first-level indicator, which represents the subjective evaluation of the quality of commercial space;

   The comprehensive evaluation module is used to obtain the comprehensive evaluation of the commercial space quality according to the objective evaluation of the commercial space quality and the subjective evaluation of the commercial space quality, and calculate the comprehensive evaluation of the commercial space quality.
10. A computer device, comprising a processor and a memory for storing a program executable by the processor, characterized in that, when the processor executes the program stored in the memory, the commercial space quality evaluation method of any one of claims 1-8 is realized .

Images