A method for multi-source real estate market data fusion and trend analysis

Abstract

The present invention relates to the field of data processing technology, and in particular to a multi-source real estate market data fusion and trend analysis method, comprising obtaining the data benchmark degree of each administrative district according to the volume of the intersecting space between data point groups of all days in different administrative districts and the GDP distribution of all administrative districts, and determining the benchmark point group of the benchmark administrative district through the data benchmark degree; adjusting the position coordinates of each data point according to the coordinates of all data points in the data point group of each administrative district, the GDP difference between each administrative district and the benchmark administrative district, and the difference of the real estate data characteristic value to obtain an adjusted marking point; clustering the adjusted marking points of all administrative districts to obtain a plurality of clusters; and performing real estate trend analysis according to the plurality of clusters. The present invention improves the accuracy of clustering results and improves the accuracy of real estate trend analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a multi-source real estate market data fusion and trend analysis method. Background Art

[0002] Multi-source real estate market data fusion and trend analysis refers to integrating real estate market data from different sources and analyzing these data to reveal market trends and patterns. This analysis can help governments, businesses and individuals make more informed decisions, including investment, planning and marketing.

[0003] In conventional technology, multiple influencing factors of an administrative district every day can be taken as a data point, and all data points can be clustered to analyze the trend of real estate based on the clustering results. Since the data point groups of all days in each administrative district are relatively concentrated, and the distance measurement between data point groups in different administrative districts is relatively large, when clustering analysis is directly performed on the data point groups of all administrative districts, the data point group of one administrative district may be directly taken as a cluster, which will lose the meaning of clustering, resulting in a reduction in the accuracy of the clustering results, thereby reducing the accuracy of the real estate trend analysis. Summary of the invention

[0004] The present invention provides a multi-source real estate market data fusion and trend analysis method to solve the existing problems.

[0005] The multi-source real estate market data fusion and trend analysis method of the present invention adopts the following technical solutions:

[0006] An embodiment of the present invention provides a multi-source real estate market data fusion and trend analysis method, the method comprising the following steps:

[0007] Get the GDP of each administrative area in any quarter and the factor series of all days, the factor series is composed of real estate factor data for each day;

[0008] Map the data in the factor sequence of each day in the sample space to obtain the data points corresponding to each day in the sample space, and the sample space is a space composed of factor data of all dimensions of real estate; obtain the data benchmark level of each administrative district according to the volume of the intersection space between each administrative district and the data point groups of all other administrative districts on all days, and the GDP distribution of all administrative districts, and determine the benchmark point group of the benchmark administrative district through the data benchmark level;

[0009] According to the coordinates of all data points in the data point group of each administrative district, the real estate data characteristic value of each administrative district is obtained. According to the difference in GDP between each administrative district and the benchmark administrative district and the difference in the real estate data characteristic value, the position adjustment coefficient of each administrative district is obtained. According to the position adjustment coefficient of each administrative district, the position coordinates of each data point are adjusted to obtain the adjusted marking points; by clustering the adjusted marking points of all administrative districts, several clusters are obtained;

[0010] The trend of multi-source real estate is analyzed based on the proportion of data points in each administrative district in each cluster and the data characteristics of real estate in each cluster.

[0011] Furthermore, the data benchmark level of each administrative district is obtained according to the volume of the intersection space between each administrative district and the data point groups of all other administrative districts on all days and the GDP distribution of all administrative districts, and the specific steps include the following:

[0012] According to the position of the GDP of each administrative district within the range of the minimum and maximum GDP of all administrative districts, the benchmark probability of each administrative district is obtained. According to the mean of the volume of the intersection space between each administrative district and all other administrative districts and the benchmark probability of each administrative district, the data benchmark degree of each administrative district is obtained.

[0013] Among them, the mean value of the volume of the intersection space between each administrative district and all other administrative districts is positively correlated with the data benchmark level of each administrative district, and the benchmark probability of each administrative district is positively correlated with the data benchmark level of each administrative district;

[0014] The specific process of obtaining the benchmark probability of each administrative district is as follows: the difference between the GDP of each administrative district and the maximum GDP of all administrative districts is recorded as the first difference of each administrative district, the difference between the GDP of each administrative district and the minimum GDP of all administrative districts is recorded as the second difference of each administrative district, the ratio of the first difference to the second difference of each administrative district is recorded as the first ratio of each administrative district, the difference between the first ratio of each administrative district and 1 is recorded as the third difference of each administrative district, and the third difference of each administrative district is negatively correlated and mapped to obtain the benchmark probability of each administrative district.

[0015] Furthermore, the step of determining the benchmark point group of the benchmark administrative area through the data benchmark level includes the following specific steps:

[0016] The administrative district with the largest data benchmark degree is selected as the benchmark administrative district, and the data point group corresponding to the benchmark administrative district is recorded as the benchmark point group.

[0017] Furthermore, the real estate data characteristic value of each administrative district is obtained according to the coordinates of all data points in the data point group of each administrative district, and the specific steps include the following:

[0018] The square sum of the coordinate values ​​of each data point in each administrative district in all dimensions is recorded as the first value of each data point in each administrative district, and the mean of the first values ​​of all data points in each administrative district is recorded as the second value of each administrative district. The real estate data characteristic value of each administrative district is obtained according to the second value of each administrative district;

[0019] Among them, the second value of each administrative district is positively correlated with the characteristic value of the real estate data of each administrative district.

[0020] Furthermore, the location adjustment coefficient of each administrative district is obtained according to the GDP difference between each administrative district and the benchmark administrative district and the difference of the characteristic value of the real estate data, and the specific steps include the following:

[0021] The GDP difference between each administrative district and the benchmark administrative district is positively correlated with the location adjustment coefficient of each administrative district, and the difference in the characteristic values ​​of real estate data between each administrative district and the benchmark administrative district is positively correlated with the location adjustment coefficient of each administrative district.

[0022] Furthermore, the position coordinates of each data point are adjusted according to the position adjustment coefficient of each administrative area to obtain the adjusted marking point, which includes the following specific steps:

[0023] Determine the adjustment direction of each administrative district data point according to the position adjustment coefficient of each administrative district, wherein the adjustment direction includes a coordinate increasing direction and a coordinate decreasing direction;

[0024] Adjusting the coordinates of the data points in each administrative district according to the position adjustment coefficient of each administrative district and the adjustment direction;

[0025] The adjusted coordinates of all data points in each administrative district are marked as the marking points of each administrative district.

[0026] Furthermore, the step of clustering the adjusted marking points of all administrative regions to obtain a number of clusters includes the following specific steps:

[0027] All the marked points in all administrative areas are clustered by agglomerative hierarchical clustering algorithm to obtain several clusters.

[0028] Furthermore, the method of assisting in analyzing the trend of multi-source real estate according to the proportion of data points of each administrative district in each cluster and the data characteristics of real estate in each cluster includes the following specific steps:

[0029] The real estate data characteristic value of each cluster is obtained according to the calculation process of the real estate data characteristic value of each administrative district; the real estate vitality level of each administrative district is obtained according to the proportion of data points of each administrative district in each cluster and the real estate data characteristic value of each cluster;

[0030] Conduct a trend analysis of real estate in each administrative district based on the degree of real estate activity in each administrative district for several consecutive quarters.

[0031] Furthermore, the real estate vitality level of each administrative district is obtained according to the proportion of data points of each administrative district in each cluster and the real estate data characteristic value of each cluster, and the specific steps include the following:

[0032] The proportion of data points of each administrative district in each cluster is recorded as the first proportion of each administrative district in each cluster. The product of the first proportion of each administrative district in each cluster and the real estate data characteristic value of each cluster is recorded as the first product value of each administrative district in each cluster. The average of the first product values ​​of each administrative district in all clusters is taken as the degree of real estate vitality of each administrative district.

[0033] Furthermore, the real estate activity level of each administrative district for several consecutive quarters is analyzed to analyze the real estate trend of each administrative district, including the following specific steps:

[0034] Through the process of obtaining the real estate vitality of each administrative district in any quarter, the real estate vitality of each administrative district for several consecutive quarters is obtained;

[0035] Based on the real estate vitality level of each administrative district for several consecutive quarters, the real estate vitality level of each administrative district in the future quarter is predicted through the autoregressive moving average model. The predicted real estate vitality level is displayed on the terminal display as a reference for real estate investment, so as to assist in analyzing the trend of multi-source real estate.

[0036] The beneficial effects of the technical solution of the present invention are as follows: the present invention obtains the data benchmark level of each administrative district according to the volume of the intersecting space between the data point groups of all days in different administrative districts and the GDP distribution of all administrative districts, determines the benchmark point group of the benchmark administrative district through the data benchmark level, and uses the benchmark point group as a reference to adjust the coordinates of the data points in the data point groups of other administrative districts; obtains the position adjustment coefficient of each administrative district according to the GDP difference between each administrative district and the benchmark administrative district and the difference in the characteristic value of real estate data, adjusts the position coordinates of each data point according to the position adjustment coefficient of each administrative district, and obtains the adjusted marking point, and by adjusting the position coordinates of the data point, the influence of the distance measurement between the data point groups of different administrative districts is reduced; by clustering the adjusted marking points of all administrative districts, a number of clusters are obtained, thereby improving the accuracy of the clustering results; according to the proportion of data points of each administrative district in each cluster and the data characteristics of the real estate of each cluster, the trend of multi-source real estate is assisted in the analysis, thereby improving the accuracy of the real estate trend analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 A flowchart of the steps of a multi-source real estate market data fusion and trend analysis method of the present invention;

[0039] Figure 2 This is a flowchart for multi-source real estate market data fusion and trend analysis. DETAILED DESCRIPTION

[0040] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the multi-source real estate market data fusion and trend analysis method proposed by the present invention, its specific implementation method, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0041] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0042] The following is a detailed description of a multi-source real estate market data fusion and trend analysis method provided by the present invention in conjunction with the accompanying drawings.

[0043] See also Figure 1 , which shows a flowchart of a multi-source real estate market data fusion and trend analysis method provided by an embodiment of the present invention, the method comprising the following steps:

[0044] Step S001: Collect data on various factors affecting real estate and GDP data.

[0045] It should be noted that in order to analyze the real estate trend of a region, it is necessary to analyze the changes in various factors affecting real estate, assisted by analyzing the trends of multi-source real estate, to comprehensively determine the development trend of real estate in a region, so as to consider whether to invest in real estate in a region.

[0046] Specifically, various factor data affecting real estate are collected every day, and data for a quarter are collected in total, so that factor data for all days of a quarter are obtained; all factor data are subjected to standard normalization processing to obtain standard normalized factor data; all factor data of each day after standard normalization are combined into a sequence, which is recorded as the daily factor sequence. Among them, various factor data include house transaction volume, transaction price, etc. Among them, the process of standard normalization processing of daily factor data is a well-known technology, and will not be described in detail here.

[0047] It should be further explained that, when analyzing the real estate market of a region, economic factors are the main factors, and the economy of a region is usually reflected by the corresponding GDP. And because the GDP of a region is generally released on a quarterly basis, the GDP of one quarter is collected separately for analysis. Therefore, it is necessary to collect the GDP of each administrative district every quarter.

[0048] At this point, we have obtained the GDP of each administrative region in each quarter and the factor series of all days.

[0049] It should be noted that since the investment situation is analyzed through the difference between real estates in different administrative regions, it is necessary to collect data from multiple administrative regions for analysis, that is, to collect the GDP of several administrative regions every quarter and the factor series of all days.

[0050] At this point, we have obtained the GDP of several administrative regions in each quarter and the factor series of all days.

[0051] Step S002: Map the data in the factor sequence of each day in the sample space to obtain the data points corresponding to each day in the sample space; obtain the data benchmark level of each administrative district based on the volume of the intersection space between each administrative district and the data point groups of all other administrative districts on all days and the GDP distribution of all administrative districts, and determine the benchmark point group of the benchmark administrative district through the data benchmark level.

[0052] It should be noted that since the differences between data point groups in different administrative districts may be large, direct clustering will lead to poor clustering results. Therefore, a central reference data point group can be selected, and the data point groups of all administrative districts can be moved closer to the center point through the relative position relationship between the reference data point group and the data point groups of other administrative districts. This can reduce the clustering differences between different administrative districts.

[0053] Specifically, by mapping the daily factor sequence collected above into a sample space, the daily factor sequence corresponds to a data point in the sample space; wherein each axis of the sample space corresponds to a factor, that is, the number of dimensions of the sample space is the same as the number of all factors.

[0054] In this embodiment, one quarter is used as an example for analysis, and any quarter is recorded as a target quarter. The data of subsequent analyses are all analyzed based on the target quarter.

[0055] The factor sequence of all days in the target quarter of each administrative district is used to determine a data point group, and several administrative districts correspond to several data point groups. Among them, one data point group corresponds to one space, and the volume of the intersection space of two data point groups is obtained by affine transformation. Among them, affine transformation is a well-known technology and will not be described in detail here.

[0056] The larger the volume of the intersecting space between the space corresponding to the data point group of each administrative district and the space corresponding to the data point group of other administrative districts, the greater the overlap between the data corresponding to each administrative district and the data corresponding to other administrative districts, that is, the greater the possibility of taking all the data corresponding to the administrative district as a reference point group; the closer the position of the GDP of each administrative district is to the middle position of the minimum and maximum value interval of the GDP of all administrative districts, the more suitable the data point group of the administrative district is to be taken as a reference point group; therefore, according to the volume of the intersecting space between the data point groups of each administrative district and all other administrative districts on all days and the GDP distribution of all administrative districts, the data reference degree of each administrative district is obtained. As an embodiment, the specific calculation method is:

[0057] According to the position of the GDP of each administrative district within the range of the minimum and maximum GDP of all administrative districts, the benchmark probability of each administrative district is obtained. According to the mean of the volume of the intersection space between each administrative district and all other administrative districts and the benchmark probability of each administrative district, the data benchmark degree of each administrative district is obtained.

[0058] Among them, the mean value of the volume of the intersection space between each administrative district and all other administrative districts is positively correlated with the data benchmark level of each administrative district, and the benchmark probability of each administrative district is positively correlated with the data benchmark level of each administrative district;

[0059] The specific process of obtaining the benchmark probability of each administrative district is as follows: the difference between the GDP of each administrative district and the maximum GDP of all administrative districts is recorded as the first difference of each administrative district, the difference between the GDP of each administrative district and the minimum GDP of all administrative districts is recorded as the second difference of each administrative district, the ratio of the first difference to the second difference of each administrative district is recorded as the first ratio of each administrative district, the difference between the first ratio of each administrative district and 1 is recorded as the third difference of each administrative district, and the third difference of each administrative district is negatively correlated and mapped to obtain the benchmark probability of each administrative district.

[0060] In one embodiment of the present invention, the data benchmark level of each administrative area is expressed by the formula:

[0061]

[0062] Where, T i,j represents the volume of the intersection space between the i-th administrative district and the j-th administrative district, n represents the number of all administrative districts, max(G) represents the maximum GDP of all administrative districts, min(G) represents the minimum GDP of all administrative districts, G i represents the GDP of the ith administrative district, max() represents the maximum value function, min() represents the minimum value function, || represents the absolute value function, exp() represents the exponential function with a natural constant as the base, Q i represents the data benchmark level of the ith administrative district, and G represents the GDP set of all administrative districts. The purpose of adding 0.01 to the denominator of the formula is to prevent the denominator from being zero.

[0063] in, It represents the mean of the volume of the intersecting space between the space corresponding to the data point group of each administrative district and the space corresponding to the data point groups of other administrative districts. The larger the mean, the greater the degree of overlap between the data corresponding to each administrative district and the data corresponding to other administrative districts, that is, the greater the possibility of treating all the data corresponding to the administrative district as a reference point group; the smaller the mean, the smaller the degree of overlap between the data corresponding to each administrative district and the data corresponding to other administrative districts, that is, the smaller the possibility of treating all the data corresponding to the administrative district as a reference point group. It represents the ratio of the GDP of each administrative district to the difference between the maximum and minimum GDP of all administrative districts. The closer the ratio is to 1, the more likely the GDP of the administrative district is in the middle, and the more likely the data point group corresponding to the administrative district is to be used as the benchmark point group; the less likely the ratio is to 1, the less likely the data point group corresponding to the administrative district is to be used as the benchmark point group.

[0064] At this point, the data benchmark level for each administrative district is obtained.

[0065] The administrative district with the largest data benchmark degree is selected as the benchmark administrative district, and the data point group corresponding to the benchmark administrative district is recorded as the benchmark point group.

[0066] At this point, the reference point group is obtained.

[0067] Step S003: According to the coordinates of all data points in the data point group of each administrative district, the real estate data characteristic value of each administrative district is obtained; according to the difference in GDP between each administrative district and the benchmark administrative district and the difference in the real estate data characteristic value, the position adjustment coefficient of each administrative district is obtained; according to the position adjustment coefficient of each administrative district, the position coordinates of each data point are adjusted to obtain adjusted marking points; by clustering the adjusted marking points of all administrative districts, a number of clusters are obtained.

[0068] It should be noted that, since each administrative district corresponds to a data point group, a real estate data feature corresponding to each data point group is obtained through all the data in the data point group, and the position of the data points in the data point group is adjusted according to the difference between the real estate data features of different data point groups.

[0069] It should be further explained that the coordinates of the data point group of each administrative district can reflect the data distribution of each administrative district. Therefore, based on the coordinates of all data points in the data point group of each administrative district, the real estate data characteristic value of each administrative district is obtained to represent the real estate characteristics of the administrative district.

[0070] As an embodiment, the specific calculation method is: record the square sum of the coordinate values ​​of each data point in each administrative district in all dimensions as the first value of each data point in each administrative district, record the mean of the first values ​​of all data points in each administrative district as the second value of each administrative district, and obtain the real estate data characteristic value of each administrative district according to the second value of each administrative district; wherein the second value of each administrative district is positively correlated with the real estate data characteristic value of each administrative district. The specific calculation formula is:

[0071]

[0072] In the formula, F i,c,v represents the data value of the cth data point in the i-th administrative district in the vth dimension, m2 represents the number of all dimensions, m1 represents the number of all data points in the data point group of each administrative district, and H i Represents the characteristic value of real estate data in the i-th administrative district.

[0073] Among them, when the data value of each data point in each dimension of each administrative district is larger, the real estate data characteristic value of each administrative district is larger, and vice versa, the real estate data characteristic value of each administrative district is smaller.

[0074] At this point, the real estate data characteristic values ​​of each administrative district are obtained.

[0075] It should be noted that since the data point group of each administrative district is relatively isolated compared to the data point groups of other administrative districts, the benchmark point group is selected and adjusted based on the relative position relationship between the benchmark point group and the data point groups of other administrative districts, so that the data points are closer. In order to achieve accurate adjustment, it is necessary to use the benchmark point group as a reference, and according to the GDP difference and the difference in the characteristic value of real estate data between the benchmark administrative district and each other administrative district in the target quarter, the position adjustment coefficient of each administrative district is obtained for subsequent position adjustment.

[0076] As an embodiment, the specific calculation method is: the GDP difference between each administrative district and the benchmark administrative district is positively correlated with the location adjustment coefficient of each administrative district, and the difference in the real estate data characteristic value between each administrative district and the benchmark administrative district is positively correlated with the location adjustment coefficient of each administrative district. In one embodiment of the present invention, the specific calculation formula is:

[0077] R i =tanh(ΔG i ×(H′-H i ))

[0078] ΔG i =norm|G′-G i |

[0079] In the formula, G i represents the GDP of the ith administrative district, G' represents the GDP of the benchmark administrative district, ΔG i represents the GDP difference between the ith administrative district and the benchmark administrative district, H' represents the characteristic value of the real estate data of the benchmark administrative district, and H i represents the characteristic value of real estate data in the i-th administrative district, R i It represents the position adjustment coefficient of the i-th administrative district, norm() represents the linear normalization function, || is the absolute value symbol, and tanh() represents the hyperbolic tangent function.

[0080] Among them, |G'-G i | represents the GDP difference between each administrative district and the benchmark administrative district. When the difference is smaller, the data point position of each administrative district needs to be adjusted less. Conversely, the data point position of each administrative district needs to be adjusted more. ΔG i ×(H'-H i ) i As (H'-H i ) weights, when the difference in the characteristic values ​​of real estate data between each administrative district and the benchmark administrative district is greater, the degree to which the data point position of each administrative district needs to be adjusted is smaller, and vice versa, the degree to which the data point position of each administrative district needs to be adjusted is greater.

[0081] At this point, the location adjustment coefficient of each administrative area is obtained.

[0082] It should be noted that in order to eliminate the clustering differences of data point groups in different administrative districts, the positions of all data points in all administrative districts are adjusted according to the differences between each administrative district and the benchmark administrative district, so as to improve the compactness between data point groups in different administrative districts; that is, the position coordinates of each data point are adjusted according to the position adjustment coefficient of each administrative district to obtain the adjusted coordinates of each data point.

[0083] As an embodiment, the specific calculation method is: determine the adjustment direction of each administrative district data point according to the position adjustment coefficient of each administrative district, and the adjustment direction includes the direction of increasing coordinates and the direction of decreasing coordinates; adjust the coordinates of the data points in each administrative district according to the position adjustment coefficient of each administrative district and the adjustment direction.

[0084] In one embodiment of the present invention, because the value range of the location adjustment coefficient of each administrative district is between -1 and 1, in this embodiment, the specific calculation formula of the location adjustment coefficient of each administrative district is:

[0085]

[0086] In the formula, R irepresents the position adjustment coefficient of the i-th administrative district, (x i,k ,y i,k ,…,z i,k ) represents the coordinates of the kth data point in the data point group of the i-th administrative district, (x i,k ,y i,k ,…,z i,k )′ represents the adjusted coordinates of the kth data point in the data point group of the ith administrative district, and || is the absolute value symbol.

[0087] At this point, the adjusted coordinates of all data points in each administrative area are obtained.

[0088] It should be noted that since multiple places may suffer the same thing in the same period, analyzing all data points in all administrative areas in the same period has important reference value and can eliminate the changes in the real estate market caused by analyzing most areas alone.

[0089] The marker points of each administrative district are obtained according to the adjusted coordinates of all data points in each administrative district, and all marker points in all administrative districts are clustered by an agglomerative hierarchical clustering algorithm to obtain several clusters. The agglomerative hierarchical clustering algorithm is a well-known technology and will not be described in detail here.

[0090] So far, several clusters are obtained.

[0091] Step S004: Assist in analyzing the trend of multi-source real estate based on the proportion of data points of each administrative district in each cluster and the data characteristics of real estate in each cluster.

[0092] It should be noted that the results of a cluster are all parts of different administrative districts with similar multidimensional data characteristics, and the distribution of all data points in each administrative district in the target quarter reflects the size characteristics of multi-source real estate data, that is, the real estate data characteristics of all data points in each administrative district in different clusters can be used to analyze the real estate vitality of each administrative district. Therefore, the real estate data characteristic value of each cluster is obtained according to the calculation process of the real estate data characteristic value of each administrative district; the real estate vitality of each administrative district is obtained according to the proportion of data points of each administrative district in each cluster and the real estate data characteristic value of each cluster.

[0093] As an embodiment, the specific calculation method is: record the proportion of data points of each administrative district in each cluster as the first proportion of each administrative district in each cluster, record the product of the first proportion of each administrative district in each cluster and the real estate data characteristic value of each cluster as the first product value of each administrative district in each cluster, and take the average of the first product values ​​of each administrative district in all clusters as the real estate vitality level of each administrative district. The specific calculation formula is:

[0094]

[0095] Where W s represents the number of all data points in the sth cluster, w i,s H represents the number of data points in the ith administrative district that appear in the sth cluster. s represents the characteristic value of the real estate data of the sth cluster, M represents the number of all clusters, P i Represents the degree of real estate vitality in the ith administrative district.

[0096] Among them, when the real estate data characteristic value of each cluster is larger, it means that the real estate house transaction volume, transaction price, etc. are higher, which means that the real estate is more dynamic; when the real estate data characteristic value of each cluster is smaller, it means that the real estate house transaction volume, transaction price, etc. are lower, which means that the real estate is less dynamic. represents the proportion of data points in each cluster in each administrative area, and the proportion is used as H s The greater the proportion, the greater the weight.

[0097] At this point, the level of real estate activity in each administrative district in the target quarter is obtained.

[0098] By obtaining the real estate vitality level of each administrative district in the target quarter, the real estate vitality level of each administrative district in several consecutive quarters can be obtained.

[0099] Based on the real estate vitality of each administrative district for several consecutive quarters, the real estate vitality of each administrative district in the future quarter is predicted through the autoregressive moving average model (ARIMA). The predicted real estate vitality is displayed on the terminal display as a reference for real estate investment, so as to assist in analyzing the trend of multi-source real estate.

[0100] The autoregressive moving average model is a well-known technology and will not be described in detail here. Figure 2 shown.

[0101] It should be noted that the exp(-x) model used in this embodiment is only used to indicate that the negative correlation and the result of the constraint model output are in the interval (0,1). In specific implementation, it can be replaced with other models with the same purpose. This embodiment only uses the exp(-x) model as an example for description without making specific limitations on it, where x refers to the input of the model.

[0102] At this point, this embodiment is completed.

[0103] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-source real estate market data fusion and trend analysis method, characterized in that: The method comprises the following steps: Get the GDP of each administrative area in any quarter and the factor series of all days, the factor series is composed of real estate factor data for each day; Map the data in the factor sequence of each day in the sample space to obtain the data points corresponding to each day in the sample space, and the sample space is a space composed of factor data of all dimensions of real estate; obtain the data benchmark level of each administrative district according to the volume of the intersection space between each administrative district and the data point groups of all other administrative districts on all days, and the GDP distribution of all administrative districts, and determine the benchmark point group of the benchmark administrative district through the data benchmark level; According to the coordinates of all data points in the data point group of each administrative district, the real estate data characteristic value of each administrative district is obtained. According to the difference in GDP between each administrative district and the benchmark administrative district and the difference in the real estate data characteristic value, the position adjustment coefficient of each administrative district is obtained. According to the position adjustment coefficient of each administrative district, the position coordinates of each data point are adjusted to obtain the adjusted marking points; by clustering the adjusted marking points of all administrative districts, several clusters are obtained; The trend of multi-source real estate is analyzed based on the proportion of data points in each administrative district in each cluster and the data characteristics of real estate in each cluster.

2. The multi-source real estate market data fusion and trend analysis method according to claim 1, characterized in that: The data benchmark level of each administrative district is obtained according to the volume of the intersection space between each administrative district and the data point groups of all other administrative districts on all days and the GDP distribution of all administrative districts, and the specific steps include the following: According to the position of the GDP of each administrative district within the range of the minimum and maximum GDP of all administrative districts, the benchmark probability of each administrative district is obtained. According to the mean of the volume of the intersection space between each administrative district and all other administrative districts and the benchmark probability of each administrative district, the data benchmark degree of each administrative district is obtained. Among them, the mean value of the volume of the intersection space between each administrative district and all other administrative districts is positively correlated with the data benchmark level of each administrative district, and the benchmark probability of each administrative district is positively correlated with the data benchmark level of each administrative district; The specific process of obtaining the benchmark probability of each administrative district is as follows: the difference between the GDP of each administrative district and the maximum GDP of all administrative districts is recorded as the first difference of each administrative district, the difference between the GDP of each administrative district and the minimum GDP of all administrative districts is recorded as the second difference of each administrative district, the ratio of the first difference to the second difference of each administrative district is recorded as the first ratio of each administrative district, the difference between the first ratio of each administrative district and 1 is recorded as the third difference of each administrative district, and the third difference of each administrative district is negatively correlated and mapped to obtain the benchmark probability of each administrative district.

3. The multi-source real estate market data fusion and trend analysis method according to claim 1 is characterized in that: The specific steps of determining the benchmark point group of the benchmark administrative area through the data benchmark level are as follows: The administrative district with the largest data benchmark degree is selected as the benchmark administrative district, and the data point group corresponding to the benchmark administrative district is recorded as the benchmark point group.

4. The multi-source real estate market data fusion and trend analysis method according to claim 1, characterized in that: The specific steps of obtaining the real estate data characteristic value of each administrative district according to the coordinates of all data points in the data point group of each administrative district are as follows: The square sum of the coordinate values ​​of each data point in each administrative district in all dimensions is recorded as the first value of each data point in each administrative district, and the mean of the first values ​​of all data points in each administrative district is recorded as the second value of each administrative district. The real estate data characteristic value of each administrative district is obtained according to the second value of each administrative district; Among them, the second value of each administrative district is positively correlated with the characteristic value of the real estate data of each administrative district.

5. The multi-source real estate market data fusion and trend analysis method according to claim 1, characterized in that: The specific steps of obtaining the position adjustment coefficient of each administrative district according to the GDP difference between each administrative district and the benchmark administrative district and the difference of the characteristic value of real estate data are as follows: The GDP difference between each administrative district and the benchmark administrative district is positively correlated with the location adjustment coefficient of each administrative district, and the difference in the characteristic values ​​of real estate data between each administrative district and the benchmark administrative district is positively correlated with the location adjustment coefficient of each administrative district.

6. The multi-source real estate market data fusion and trend analysis method according to claim 1, characterized in that: The specific steps of adjusting the position coordinates of each data point according to the position adjustment coefficient of each administrative district to obtain the adjusted marking point are as follows: Determine the adjustment direction of each administrative district data point according to the position adjustment coefficient of each administrative district, wherein the adjustment direction includes a coordinate increasing direction and a coordinate decreasing direction; Adjusting the coordinates of the data points in each administrative district according to the position adjustment coefficient of each administrative district and the adjustment direction; The adjusted coordinates of all data points in each administrative district are marked as the marking points of each administrative district.

7. The multi-source real estate market data fusion and trend analysis method according to claim 1, characterized in that: The method of clustering the adjusted marked points of all administrative regions to obtain several clusters includes the following specific steps: All the marked points in all administrative areas are clustered by agglomerative hierarchical clustering algorithm to obtain several clusters.

8. The multi-source real estate market data fusion and trend analysis method according to claim 1, characterized in that: The method of assisting in analyzing the trend of multi-source real estate based on the proportion of data points of each administrative district in each cluster and the data characteristics of real estate in each cluster includes the following specific steps: The real estate data characteristic value of each cluster is obtained according to the calculation process of the real estate data characteristic value of each administrative district; the real estate vitality level of each administrative district is obtained according to the proportion of data points of each administrative district in each cluster and the real estate data characteristic value of each cluster; Conduct a trend analysis of real estate in each administrative district based on the degree of real estate activity in each administrative district for several consecutive quarters.

9. The multi-source real estate market data fusion and trend analysis method according to claim 8, characterized in that: The specific steps of obtaining the degree of real estate vitality of each administrative district according to the proportion of data points of each administrative district in each cluster and the real estate data characteristic value of each cluster are as follows: The proportion of data points of each administrative district in each cluster is recorded as the first proportion of each administrative district in each cluster. The product of the first proportion of each administrative district in each cluster and the real estate data characteristic value of each cluster is recorded as the first product value of each administrative district in each cluster. The average of the first product values ​​of each administrative district in all clusters is taken as the degree of real estate vitality of each administrative district.

10. The multi-source real estate market data fusion and trend analysis method according to claim 8, characterized in that: The specific steps of conducting trend analysis on real estate in each administrative district through the degree of real estate vitality in each administrative district for several consecutive quarters are as follows: Through the process of obtaining the real estate vitality of each administrative district in any quarter, the real estate vitality of each administrative district for several consecutive quarters is obtained; Based on the real estate vitality level of each administrative district for several consecutive quarters, the real estate vitality level of each administrative district in the future quarter is predicted through the autoregressive moving average model. The predicted real estate vitality level is displayed on the terminal display as a reference for real estate investment, so as to assist in analyzing the trend of multi-source real estate.

Images