Urban vitality evaluation system based on multi-source dynamic big data fusion analysis

Abstract

The present invention relates to the field of urban vitality evaluation, and specifically to an urban vitality evaluation system based on multi-source dynamic big data fusion analysis. The system first obtains vitality data of different dimensions of the city to be tested every day, as well as meteorological data of different meteorological types, and obtains daily exercise suitability based on the meteorological data of different meteorological types and their change characteristics. The correlation between the vitality data of the target dimension of each day and the exercise suitability of each day is analyzed to obtain the degree of resistance to weather interference of the target dimension. According to the change of the vitality data of the target dimension, the correlation between the vitality data between the target dimension and other dimensions, and the degree of resistance to weather interference of each dimension, the vitality evaluation weight of the target dimension is obtained, and the vitality data of all dimensions are reduced in dimension. Based on the data after dimension reduction, the vitality of the city to be tested is evaluated. The present invention can effectively reduce the dimension of multi-dimensional vitality data and improve the effect of urban vitality evaluation.

Description

Technical Field

[0001] The present invention relates to the field of urban vitality evaluation, and in particular to an urban vitality evaluation system based on multi-source dynamic big data fusion analysis. Background Art

[0002] The urban vitality evaluation system is mainly used to evaluate the living conditions of urban residents. It not only provides a scientific basis for urban planning and management, but also has important significance for improving the quality of life of residents and promoting economic development.

[0003] Usually, it is necessary to evaluate urban vitality by integrating multi-source data, that is, data of multiple different dimensions. Since there are some redundant dimensions with low correlation with urban vitality evaluation in multiple dimensions, in related technologies, multidimensional data are usually subjected to dimensionality reduction processing, and the reduced dimensionality data are used to evaluate urban vitality. However, the existing dimensionality reduction method only processes the data according to the statistical distribution characteristics, and does not consider the different importance of data of different dimensions to the evaluation of urban vitality. As a result, the existing dimensionality reduction processing will result in the loss of important dimensional data and the retention of unimportant dimensional data, which reduces the effect of urban vitality evaluation. Summary of the Invention

[0004] In order to solve the technical problem that the existing dimensionality reduction process results in the loss of important dimensional data and the retention of unimportant dimensional data, thereby reducing the effectiveness of urban vitality evaluation, the present invention aims to provide an urban vitality evaluation system based on multi-source dynamic big data fusion analysis. The technical solution adopted is as follows:

[0005] The present invention also proposes an urban vitality evaluation system based on multi-source dynamic big data fusion analysis, the system comprising:

[0006] The data collection module is used to obtain the vitality data of different dimensions of the city under test every day within a preset time period, and also obtain the meteorological data of different meteorological types every day;

[0007] The weather impact analysis module is used to obtain daily weather indicators based on meteorological data of different weather types each day; obtain daily exercise suitability based on the daily weather indicators and the differences between the weather indicators of each day and adjacent days; and analyze the correlation between the vitality data of the target dimension and the exercise suitability of each day within a preset time period, taking any dimension as the target dimension, to obtain the degree of resistance to weather interference of the target dimension;

[0008] A data dimensionality reduction module is configured to obtain a vitality evaluation weight for a target dimension based on changes in the vitality data of the target dimension on each day and the degree of resistance to weather interference of the target dimension, the correlation between the vitality data of the target dimension on each day and other dimensions other than the target dimension, and the degree of resistance to weather interference of other dimensions other than the target dimension; perform dimensionality reduction processing on the vitality data of all dimensions based on the vitality evaluation weight of each dimension to obtain reduced-dimensionality data;

[0009] The city vitality evaluation module is used to evaluate the vitality of the city under test based on the data after dimensionality reduction.

[0010] Furthermore, obtaining daily weather indicators includes:

[0011] Take any day as the target day, and take the absolute value of the difference between the meteorological data of each meteorological type on the target day and the standard suitable data of each meteorological type as the meteorological deviation degree of each meteorological type on the target day;

[0012] The weather deviation degrees of all weather types on the target day are integrated and normalized to obtain the weather index of the target day.

[0013] Furthermore, obtaining daily exercise fitness includes:

[0014] The absolute value of the difference between the weather index on the target day and the previous day is used as the weather index change on the target day;

[0015] The weather index and the change amount of the weather index on the target day are integrated and negatively correlated to obtain the exercise suitability on the target day.

[0016] Furthermore, obtaining the degree of weather resistance of the target dimension includes:

[0017] Performing a fitting process on the exercise suitability of each day in a preset time period to obtain a plurality of time segments of the preset time period;

[0018] Take any time segment as the target time segment, and sort the vitality data of the target dimension in each day of the target time segment in time sequence as the first vitality data sequence of the target dimension in the target time segment; and take the exercise fitness sequence of each day of the target time segment in time sequence as the exercise fitness sequence of the target time segment;

[0019] Performing negative correlation mapping on the absolute values ​​of the Pearson correlation coefficients between the first vitality data sequence and the exercise suitability sequence to obtain weather interference resistance parameters of the target dimension in the target time segment;

[0020] The two time segments closest to the target time segment are used as reference time segments. The difference in length between the target time segment and each reference time segment is used as the relative length difference between the target time segment and each reference time segment. The average of the relative length differences between the target time segment and all reference time segments is used as the numerator, the length of the target time segment is used as the denominator, and the value after normalization of the ratio is used as the weight parameter of the target dimension in the target time segment.

[0021] Based on the weight parameter of the target dimension in each time segment, the anti-weather interference parameter of the target dimension in each time segment is weighted and summed to obtain the anti-weather interference degree of the target dimension.

[0022] Furthermore, performing fitting processing on the exercise suitability of each day in the preset time period to obtain a plurality of time segments of the preset time period includes:

[0023] Using the least square method, curve fitting is performed on the two-dimensional data points constituted by the daily exercise fitness to obtain a fitting curve;

[0024] The extreme points on the fitting curve are used as segmentation points, and the preset time period is divided into multiple time segments using each segmentation point.

[0025] Furthermore, obtaining the vitality evaluation weight of the target dimension includes:

[0026] Based on the acquisition method of multiple time segments of a preset time period, the vitality data of the target dimension in each day of the preset time period is fitted, and the preset time period is divided to obtain multiple time sub-segments of the preset time period with respect to the target dimension, as well as the fitted vitality data of the target dimension in each day of the preset time period;

[0027] Take any time subsegment as the target time subsegment, take the absolute value of the difference between the vitality data of the target dimension on each day in the target time subsegment and the next adjacent day as the change in vitality data of the target dimension on each day in the target time subsegment, and take the product of the average value of the change in vitality data of the target dimension on all days in the target time subsegment and the weather interference resistance of the target dimension as the first trend change parameter of the target dimension in the target time subsegment;

[0028] Dimensions other than the target dimension are used as reference dimensions, correlation between the target dimension and the vitality data of each reference dimension on each day in the target time subsegment is analyzed, and combined with the weather interference resistance level of each reference dimension, a second trend change parameter of the target dimension in the target time subsegment is obtained;

[0029] The first trend change parameter and the second trend change parameter are combined to obtain a comprehensive trend change parameter of the target dimension in the target time subsegment, and the cumulative value of the comprehensive trend change parameter of the target dimension in all time subsegments is used as the trend change degree of the target dimension;

[0030] Obtaining the interference possibility of the target dimension according to the difference between the vitality data and the fitted vitality data of the target dimension in each day of a preset time period;

[0031] The degree of trend change is taken as the numerator, the possibility of interference is taken as the denominator, and the comparison value is normalized to obtain the vitality evaluation weight of the target dimension.

[0032] Furthermore, obtaining a second trend change parameter of the target dimension in the target time subsegment includes:

[0033] Take any reference dimension as the target reference dimension, sort the vitality data of the target dimension in each day of the target time subsegment in time sequence as the second vitality data sequence of the target dimension in the target time subsegment, and take the vitality data of the target reference dimension in each day of the target time subsegment in time sequence as the third vitality data sequence of the target reference dimension in the target time subsegment;

[0034] The product of the absolute value of the Pearson correlation coefficient between the second vitality data sequence and the third vitality data sequence and the weather interference resistance level of the target reference dimension is used as the data change correlation between the target dimension and the target reference dimension in the target time subsegment;

[0035] The accumulated value of the data change correlation between the target dimension and all reference dimensions in the target time subsegment is used as the second trend change parameter of the target dimension in the target time subsegment.

[0036] Furthermore, obtaining the interference possibility of the target dimension includes:

[0037] The absolute value of the difference between the vitality data and the fitted vitality data of the target dimension in each day of the preset time period is used as the fitting error of the target dimension in each day;

[0038] The average value of the fitting errors of the target dimension in all days within a preset time period is used as the interference possibility of the target dimension.

[0039] Furthermore, the data after dimension reduction is obtained includes:

[0040] The vitality data of each dimension and the vitality evaluation weight are input into a weighted PCA algorithm, and the weighted PCA algorithm is used to perform dimensionality reduction processing to obtain data after dimensionality reduction.

[0041] Furthermore, the vitality evaluation of the city to be tested includes:

[0042] The urban vitality data set was searched on the Internet, and each urban vitality data was scored and labeled by manual scoring. A neural network was trained on the labeled urban vitality data, and the dimensionality-reduced data was input into the trained neural network to obtain the vitality score of the tested city.

[0043] The present invention has the following beneficial effects:

[0044] The present invention takes into account that after the existing dimensionality reduction processing, important dimensional data may be lost and unimportant dimensional data may be retained, thereby reducing the effect of urban vitality evaluation, because firstly, the vitality data of different dimensions of the city to be tested are obtained every day within a preset time period, and considering that the urban vitality evaluation is easily affected by the weather, the meteorological data of different meteorological types are also obtained every day. Since the key to evaluating the level of urban vitality lies in the quality of the city's fitness trend, and the fitness type reflected by the vitality data of different dimensions is easily affected by the weather, thus affecting the subsequent evaluation of urban vitality, the quality of the daily weather can be reflected first by weather indicators. Since the better the weather every day and the less obvious the weather changes, the more suitable it is for fitness exercises, the degree to which the daily weather conditions are suitable for fitness exercises can be reflected by exercise suitability. Taking into account The types of fitness exercises reflected by different dimensions are affected by the weather in different ways. Therefore, the degree of resistance to weather interference is used to reflect the degree to which the fitness exercise types represented by the target dimension are interfered with by the weather, providing a basis for subsequent data dimensionality reduction. Since the vitality data of different dimensions are of different importance to the evaluation of urban vitality, the more obvious the change characteristics of the vitality data of the target dimension and the stronger the correlation between the target dimension and other dimensions, the more important the target dimension is. At the same time, the degree of resistance to weather interference of each dimension is added to the analysis process to improve the accuracy of the analysis. The importance of the target dimension to the evaluation of urban vitality is reflected by the obtained vitality evaluation weight. Then, in the dimensionality reduction process, the vitality evaluation weight of each dimension is added to improve the effect of data dimensionality reduction and retain the data of important dimensions as much as possible. Then, through the data after dimensionality reduction, a more effective vitality evaluation of the tested city is carried out. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. 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 any creative work.

[0046] Figure 1 A block diagram of an urban vitality evaluation system based on multi-source dynamic big data fusion analysis provided by one embodiment of the present invention;

[0047] Figure 2 A flow chart of a method for obtaining the degree of weather resistance of a target dimension provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features and effects of an urban vitality evaluation system based on multi-source dynamic big data fusion analysis proposed by the present invention. 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 of one or more embodiments may be combined in any suitable form.

[0049] 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.

[0050] The following describes in detail a specific solution of an urban vitality evaluation system based on multi-source dynamic big data fusion analysis provided by the present invention with reference to the accompanying drawings.

[0051] See also Figure 1 , which shows a block diagram of a city vitality evaluation system based on multi-source dynamic big data fusion analysis provided by an embodiment of the present invention. The system includes: a data acquisition module 101, a weather impact analysis module 102, a data dimensionality reduction module 103, and a city vitality evaluation module 104.

[0052] The data collection module 101 is used to obtain the vitality data of different dimensions of the city to be tested every day within a preset time period, and to obtain the meteorological data of different meteorological types every day.

[0053] The evaluation of urban vitality includes many aspects, such as the construction and development of the city in national fitness, the health status and lifestyle of urban residents, etc. Therefore, it is necessary to combine multi-source and multi-dimensional data such as sports infrastructure construction, national fitness participation, diversity of sports activities, and residents' health status to evaluate urban vitality.

[0054] The embodiment of the present invention first collects vitality data of different dimensions of the city to be tested every day within a preset time period through the website or database of the Sports Bureau, media or Internet platforms, etc., wherein the vitality data of different dimensions include, for example, public health data such as the obesity rate or chronic diseases of urban residents, fitness consumption data such as urban residents' fitness expenditure and sports product shopping consumption data, sports activity and facility data such as public fitness area coverage rate, number of sports parks, number of participants in sports activities and per capita stadium area, as well as digital platform data such as the number of sports app downloads and the number of participants in sports-related topics on social media.

[0055] It should be noted that since the dimensions of vitality data in different dimensions are different, in order to improve the accuracy of subsequent data analysis and processing, it is also necessary to perform standardized preprocessing on the collected vitality data of each dimension to eliminate the influence of the dimension. The standardized preprocessing of data is a technical means well known to technicians in this field and will not be elaborated here.

[0056] At the same time, considering that some types of fitness exercises related to urban vitality evaluation, such as outdoor fitness exercises, are easily affected by the weather, in order to improve the effect of subsequent urban vitality evaluation, the embodiment of the present invention also needs to collect meteorological data of different meteorological types every day in a preset time period in the city to be tested through meteorological websites or meteorological service APIs, where different meteorological types include, for example, temperature, humidity, wind speed, etc.

[0057] The preset time period is set to 1 year, and the specific value of the preset time period can also be set by the implementer according to the specific implementation scenario, and is not limited here.

[0058] The weather impact analysis module 102 is used to obtain daily weather indicators based on the meteorological data of different weather types in each day; obtain daily exercise suitability based on the daily weather indicators and the differences in weather indicators between each day and adjacent days; take any dimension as the target dimension, analyze the correlation between the vitality data of the target dimension of each day within a preset time period and the exercise suitability of each day, and obtain the degree of resistance to weather interference of the target dimension.

[0059] Since the key to evaluating the level of urban vitality lies in the fitness trends of the city's residents, and the types of fitness reflected by vitality data in different dimensions are easily affected by weather, thus affecting the subsequent evaluation of urban vitality, for example, when the weather is bad, the number of urban residents participating in fitness is relatively small. Therefore, we can first obtain daily weather indicators based on meteorological data of different weather types every day, and reflect the daily weather conditions through weather indicators. Subsequently, based on the daily weather indicators and the changes in weather indicators of adjacent days, we can accurately analyze the suitability of daily weather conditions for urban residents' fitness exercises.

[0060] Preferably, in one embodiment of the present invention, the method for obtaining daily weather indicators specifically includes:

[0061] Take any day as the target day. Since each meteorological type has a standard suitable data that is most suitable for exercise, for example, the temperature that is most suitable for exercise is usually 25 degrees Celsius. The closer the meteorological data of a certain meteorological type on the target day is to the standard suitable data of that meteorological type, the better the weather conditions on the target day are, and the more suitable it is for urban residents to exercise. Therefore, the absolute value of the difference between the meteorological data of each meteorological type on the target day and the standard suitable data of each meteorological type can be used as the meteorological deviation degree of each meteorological type on the target day. The smaller the meteorological deviation degree, the closer the meteorological data of each meteorological type on the target day is to the standard suitable data.

[0062] Then, the degree of meteorological deviation of all meteorological types on the target day can be integrated and normalized to obtain the weather index of the target day. The smaller the weather index, the better the weather conditions on the target day and the more suitable it is for urban residents to exercise.

[0063] In an embodiment of the present invention, the meteorological deviation degrees of all meteorological types in the target day may be integrated by calculating the sum or product of the meteorological deviation degrees of all meteorological types in the target day, which is not limited here.

[0064] In one embodiment of the present invention, the normalization processing can be specifically, for example, maximum and minimum value normalization processing, and the normalization in subsequent steps can all adopt maximum and minimum value normalization processing. In other embodiments of the present invention, other normalization methods can be selected according to the specific range of numerical values, which will not be repeated here.

[0065] As an example, in one embodiment of the present invention, the expression of the weather index of the target day may be specifically, for example:

[0066]

[0067] Where A represents the weather index of the target day; b i represents the meteorological data of the i-th meteorological type on the target day; b i ′ represents the standard suitable data of the i-th meteorological type, which is a known value; |b i -b i ′ | represents the degree of meteorological deviation of the i-th meteorological type on the target day; I represents the number of meteorological types; norm() represents the normalization function.

[0068] The same method as above can be used to obtain the daily weather index. The smaller the weather index of a certain day and the more stable the change in the weather index between that day and the adjacent days, the more stable the weather conditions on that day and the more suitable it is for urban residents to exercise. Therefore, the daily exercise suitability can be obtained based on the daily weather index and the difference in weather index between each day and the adjacent days. The exercise suitability reflects the degree to which the daily weather conditions are suitable for fitness exercises. The larger the exercise suitability of a certain day, the more suitable the weather conditions on that day are for urban residents to exercise, and the higher the enthusiasm of urban residents for fitness exercises, which provides a data basis for the subsequent calculation and analysis of the degree of resistance to weather interference in each dimension.

[0069] Preferably, in one embodiment of the present invention, the method for obtaining daily exercise fitness specifically includes:

[0070] The absolute value of the difference between the weather index of the target day and the adjacent previous day is taken as the weather index change of the target day. The smaller the weather index change, the more stable the weather conditions on the target day are relative to the adjacent previous day.

[0071] It should be noted that, since there is no adjacent previous day to the first day, the average of the weather index changes of all days after the first day can be used as the weather index change of the first day to facilitate smooth subsequent calculations.

[0072] The smaller the weather index of the target day and the smaller the change in the weather index, the more suitable the weather conditions on the target day are for urban residents to exercise. Therefore, the weather index and the change in the weather index on the target day can be integrated and negatively correlated to obtain the exercise suitability of the target day.

[0073] In the embodiment of the present invention, the weather index of the target day and the weather index change amount can be combined by calculating the sum or product of the two, which is not limited here.

[0074] As an example, in one embodiment of the present invention, the expression of exercise suitability on the target day may be specifically, for example, as follows:

[0075]

[0076] Wherein, B represents the exercise suitability on the target day; A represents the weather index on the target day; ΔA represents the change in the weather index on the target day; ε1 represents the preset first adjustment parameter, which is used to prevent the denominator from being 0. The value range of ε1 is [0.001, 0.01]. In one embodiment of the present invention, ε1 is set to 0.01. The specific value of ε1 can also be set by the implementer according to the specific implementation scenario and is not limited here.

[0077] It should be noted that in other embodiments of the present invention, negative correlation mapping may be achieved through other basic mathematical operations, which will not be described in detail here.

[0078] The same method described above can be used to obtain daily exercise suitability. To perform a fusion analysis of vitality data from different dimensions, it is necessary to evaluate the urban vitality information reflected by the vitality data from different dimensions. Since the types of fitness exercises reflected by different dimensions are affected by weather conditions to varying degrees, for example, indoor exercises related to gyms, indoor basketball courts, and indoor badminton courts are less affected by weather conditions, while outdoor exercises are more affected by weather conditions. The dimensions corresponding to exercises less affected by weather conditions may be misjudged as carrying more information due to their different data change characteristics from other dimensions. Therefore, embodiments of the present invention require analyzing the impact of weather conditions on data changes in different dimensions. First, any dimension is used as a target dimension. The lower the correlation between the vitality data of the target dimension and the exercise suitability of each day, the less affected the fitness exercises reflected by the target dimension are by weather conditions. Therefore, the correlation between the vitality data of the target dimension and the exercise suitability of each day within a preset time period can be analyzed to obtain the target dimension's resistance to weather interference. The greater the resistance to weather interference, the less affected the fitness exercises reflected by the target dimension are by weather conditions, providing a basis for subsequent data dimensionality reduction.

[0079] Preferably, in one embodiment of the present invention, the method for obtaining the weather interference resistance degree of the target dimension specifically includes:

[0080] See also Figure 2 , which shows a flow chart of a method for obtaining the degree of weather resistance of a target dimension provided by an embodiment of the present invention.

[0081] Step S201: Fitting the exercise suitability of each day in a preset time period to obtain multiple time segments of the preset time period.

[0082] Since the changing trend of exercise suitability rises and falls during the entire preset time period, it is not conducive to the subsequent correlation analysis. Therefore, in order to improve the accuracy of the subsequent calculation and analysis of the degree of resistance to weather interference in the target dimension, the exercise suitability of each day in the preset time period is first fitted to obtain multiple time segments of the preset time period, so that the changing trend of exercise suitability in a specific time segment is consistent. That is to say, within a specific time segment, the change of exercise suitability of each day is monotonic.

[0083] Preferably, in one embodiment of the present invention, the method for obtaining multiple time segments of a preset time period specifically includes:

[0084] Using the existing least squares method, curve fitting is performed on the two-dimensional data points consisting of each day and each day's exercise suitability to obtain a fitting curve. In other embodiments of the present invention, other curve fitting methods may also be used, which are not limited here.

[0085] The extreme points on the fitting curve are used as segmentation points, and the preset time period is divided into multiple time segments using each segmentation point. The extreme points include maximum points and minimum points. The extreme points can be obtained through the existing Newton method, which will not be described in detail here.

[0086] Step S202: Take any time segment as the target time segment, and use the sequence formed by sorting the vitality data of the target dimension of each day in the target time segment in chronological order as the first vitality data sequence of the target dimension in the target time segment; use the sequence formed by sorting the exercise suitability of each day in the target time segment in chronological order as the exercise suitability sequence of the target time segment; perform negative correlation mapping on the absolute value of the Pearson correlation coefficient between the first vitality data sequence and the exercise suitability sequence to obtain the anti-weather interference parameter of the target dimension in the target time segment.

[0087] From the above analysis, it can be seen that the lower the correlation between the vitality data of the target dimension on each day and the exercise suitability on each day, the weaker the fitness exercise reflected by the target dimension is affected by weather conditions. Therefore, after the preset time period is divided into multiple time segments, the correlation between the vitality data and exercise suitability of the target dimension in each time segment can be analyzed. First, any time segment is taken as the target time segment, and the sequence formed by sorting the vitality data of the target dimension in each day of the target time segment in time series is used as the first vitality data sequence of the target dimension in the target time segment; the sequence formed by sorting the exercise suitability of each day of the target time segment in time series is used as the exercise suitability sequence of the target time segment, and then the absolute value of the Pearson correlation coefficient between the first vitality data sequence and the exercise suitability sequence is negatively correlated to obtain the anti-weather interference parameter of the target dimension in the target time segment. The larger the anti-weather interference parameter, the less the fitness exercise reflected by the target dimension is affected by weather conditions in the target time segment. Subsequently, the anti-weather interference parameter of the target dimension in each time segment can be used as the data basis to accurately calculate the anti-weather interference degree of the target dimension.

[0088] As an example, in one embodiment of the present invention, the expression of the weather interference resistance parameter of the target dimension in the target time segment can be specifically, for example:

[0089]

[0090] Among them, C represents the anti-weather interference parameter of the target dimension in the target time segment; ρ represents the Pearson correlation coefficient between the first vitality data sequence of the target dimension in the target time segment and the exercise suitability sequence of the target time segment; ε2 represents the preset second adjustment parameter, which is used to prevent the denominator from being 0. The value range of ε2 is [0.001, 0.01]. In one embodiment of the present invention, ε2 is set to 0.01. The specific value of ε2 can also be set by the implementer according to the specific implementation scenario, and is not limited here.

[0091] The anti-weather interference parameters of the target dimension in each time segment can be obtained by the same method as above.

[0092] Step S203: The two time segments closest to the target time segment are used as reference time segments, the difference in length between the target time segment and each reference time segment is used as the relative length difference between the target time segment and each reference time segment, the average of the relative length difference between the target time segment and all reference time segments is used as the numerator, the length of the target time segment is used as the denominator, and the value after normalization of the ratio is used as the weight parameter of the target dimension in the target time segment.

[0093] Since different time segments represent a monotonic change process of exercise fitness, and the fluctuation characteristics of exercise fitness in different time segments are also different, the influence of exercise fitness on activities such as indoor fitness in time segments with a shorter duration of change is relatively weak, while for activities such as outdoor fitness that are greatly affected by weather conditions, even a shorter period of time will still have an impact on fitness. Therefore, in order to improve the calculation accuracy of the degree of resistance to weather interference of the target dimension, the two time segments closest to the target time segment are first used as reference time segments, and the difference in length between the target time segment and each reference time segment is used as the target time segment. The relative length difference between the target time segment and each reference time segment, the shorter the length of the target time segment and the relatively longer the length of the reference time segment, the more accurate the analysis of the degree to which the fitness exercise reflected by the target dimension is affected by weather conditions within the target time segment. Therefore, the average value of the relative length difference between the target time segment and all reference time segments can be used as the numerator, the length of the target time segment can be used as the denominator, and the value after normalization of the ratio can be used as the weight parameter of the target dimension in the target time segment. The weight parameter can be used later to make weighted adjustments to the weather interference resistance parameter to improve the calculation accuracy of the weather interference resistance degree of the target dimension.

[0094] It should be noted that for the first time segment, the two time segments closest to it are the second time segment and the third time segment; for the last time segment, the two time segments closest to it are the first-to-last time segment and the second-to-last time segment; for any other time segment except the first and last time segments, the two time segments closest to it are the two adjacent time segments. At the same time, the length of a time segment can be expressed by the number of days contained in the time segment.

[0095] As an example, in one embodiment of the present invention, the expression of the weight parameter of the target dimension in the target time segment can be specifically, for example:

[0096]

[0097] Where W represents the weight parameter of the target dimension in the target time segment; L represents the length of the target time segment, L≠0; L ′ Indicates the length of the first reference time segment of the target time segment; L ′ -L represents the relative length difference between the target time segment and the first reference time segment; L″ represents the length of the second reference time segment of the target time segment; L″-L represents the relative length difference between the target time segment and the second reference time segment; sigmoid() represents the activation function, which is used for normalization processing.

[0098] The weight parameter of the target dimension in each time segment can be obtained by the same method as above.

[0099] Step S204: Based on the weight parameter of the target dimension in each time segment, the anti-weather interference parameters of the target dimension in each time segment are weighted and summed to obtain the anti-weather interference degree of the target dimension.

[0100] Then, the weather resistance parameters of the target dimension in each time segment can be combined, and the weight parameters of the target dimension in each time segment can be used for weighted adjustment to obtain the weather resistance degree of the target dimension.

[0101] As an example, in one embodiment of the present invention, the expression of the weather interference resistance degree of the target dimension can be specifically, for example, as follows:

[0102]

[0103] Where D represents the degree of weather resistance of the target dimension; W j Represents the weight parameter of the target dimension in the jth time segment; C j represents the anti-weather interference parameter of the target dimension in the jth time segment; J represents the number of time segments in the preset time period.

[0104] The same method as above can be used to obtain the degree of resistance to weather interference in each dimension.

[0105] The data dimensionality reduction module 103 is used to obtain the vitality evaluation weight of the target dimension based on the changes in the vitality data of the target dimension on each day and the degree of resistance to weather interference of the target dimension, the correlation between the vitality data of the target dimension on each day and other dimensions except the target dimension, and the degree of resistance to weather interference of other dimensions except the target dimension; according to the vitality evaluation weight of each dimension, the vitality data of all dimensions are reduced in dimension to obtain the reduced data.

[0106] Since the importance of vitality data of different dimensions to the evaluation of urban vitality is different, it is necessary to analyze the historical dynamic change characteristics of vitality data of each dimension. In the context of the current mobile network social platform, the trend of sports type changes and the speed of dissemination are relatively fast. When evaluating urban vitality, it is necessary to magnify the importance of dimensions that can characterize the trend changes of sports type in the evaluation process. Vitality data of this dimension can carry more information about urban vitality. The trend of sports type in different periods needs to be reflected through data of multiple dimensions. When the sports type has a certain trend of change, the dimension that reflects the original sports type may usually change. For example, when badminton gradually becomes popular on the online social platform, the badminton hall data will show a certain degree of growth. After a period of time, when table tennis becomes popular, the original badminton hall data may decline, while the table tennis hall and table tennis related shopping data will show a certain growth. Therefore, it can be Combined with the correlation between the vitality data of different dimensions, the analysis is conducted to evaluate the degree of expression of the dynamic movement trend of data of different dimensions. That is to say, the more obvious the change characteristics of the vitality data of the target dimension and the stronger the correlation between the target dimension and other dimensions, the more important the target dimension is. Therefore, the changes in the vitality data of the target dimension on each day and the correlation of the vitality data between the target dimension on each day and other dimensions except the target dimension can be analyzed. At the same time, the degree of resistance to weather interference of each dimension is added to the analysis process to improve the accuracy of the analysis, and the importance of the target dimension to the evaluation of urban vitality is reflected by the obtained vitality evaluation weight. The larger the vitality evaluation weight, the more important the target dimension is in the urban vitality evaluation. Subsequently, based on the vitality evaluation weight of each dimension, more effective dimensionality reduction processing can be achieved, and the vitality data of the dimensions that are more important to the urban vitality evaluation can be retained to improve the effect of urban vitality evaluation.

[0107] Preferably, in one embodiment of the present invention, the method for obtaining the weather interference resistance degree of the target dimension specifically includes:

[0108] First, based on the acquisition method of multiple time segments of a preset time period, the vitality data of the target dimension in each day of the preset time period is fitted, and the preset time period is divided to obtain multiple time sub-segments of the target dimension in the preset time period, so that the change of the vitality data of the target dimension in a specific time sub-segment is monotonic, thereby improving the accuracy of subsequent calculation and analysis. At the same time, in the fitting process, the fitted vitality data of the target dimension in each day of the preset time period is obtained, wherein the specific acquisition process of the multiple time sub-segments of the target dimension in the preset time period is as follows: using the least squares method or other curve fitting methods, curve fitting is performed on the two-dimensional data points constituted by the vitality data of the target dimension of each day and each day to obtain a fitting curve, and the extreme value points on the fitting curve are used as segmentation points, and each segmentation point is used to divide the preset time period into multiple time sub-segments of the target dimension.

[0109] Then, any time sub-segment is taken as the target time sub-segment, and the absolute value of the difference in the vitality data of the target dimension between each day in the target time sub-segment and the adjacent next day is taken as the change in the vitality data of the target dimension for each day in the target time sub-segment. The larger the change in the vitality data, the more obvious the change characteristics of the vitality data of the target dimension for each day in the target time sub-segment. At the same time, the influence of weather conditions on the target dimension is added to the calculation and analysis process, and the product value of the average value of the change in the vitality data of the target dimension for all days in the target time sub-segment and the degree of resistance to weather interference of the target dimension is taken as the first trend change parameter of the target dimension in the target time sub-segment. The larger the first trend change parameter, the more obvious the data change of the target dimension in the target time sub-segment.

[0110] It should be noted that if there is no adjacent next day to the last day of the target time subsegment, the average of the vitality data changes of all days before the last day of the target time subsegment can be used as the vitality data change of the last day.

[0111] As an example, in one embodiment of the present invention, the expression of the first trend change parameter of the target dimension in the target time subsegment may be specifically, for example, as follows:

[0112]

[0113] Among them, E1 represents the first trend change parameter of the target dimension in the target time sub-segment; D represents the degree of weather interference resistance of the target dimension; Δa n Indicates the change in vitality data of the target dimension on the nth day in the target time subsegment; N represents the number of days in the target time subsegment.

[0114] Taking dimensions other than the target dimension as reference dimensions, the correlation between the target dimension and the vitality data of each reference dimension on each day in the target time sub-segment is analyzed. Combined with the degree of resistance to weather interference of each reference dimension, the second trend change parameter of the target dimension in the target time sub-segment is obtained. The larger the second trend change parameter, the more obvious the data change of the target dimension in the target time sub-segment.

[0115] Preferably, in one embodiment of the present invention, the method for obtaining the second trend change parameter of the target dimension in the target time subsegment specifically includes:

[0116] Take any reference dimension as the target reference dimension, and use the sequence formed by sorting the vitality data of the target dimension in each day of the target time sub-segment in chronological order as the second vitality data sequence of the target dimension in the target time sub-segment, and use the sequence formed by sorting the vitality data of the target reference dimension in each day of the target time sub-segment in chronological order as the third vitality data sequence of the target reference dimension in the target time sub-segment. At the same time, the influence of weather conditions on the target reference dimension is added to the correlation analysis process, and the product value of the absolute value of the Pearson correlation coefficient between the second vitality data sequence and the third vitality data sequence and the degree of resistance to weather interference of the target reference dimension is used as the data change correlation between the target dimension and the target reference dimension in the target time sub-segment. The greater the data change correlation, the more correlated the changes in the vitality data between the target dimension and the target reference dimension in the target time sub-segment are.

[0117] The same method as above can be used to obtain the data change correlation between the target dimension and each reference dimension in the target time sub-segment, and then the accumulated value of the data change correlation between the target dimension and all reference dimensions in the target time sub-segment can be used as the second trend change parameter of the target dimension in the target time sub-segment.

[0118] As an example, in one embodiment of the present invention, the expression of the second trend change parameter of the target dimension in the target time subsegment can be specifically, for example, as follows:

[0119]

[0120] Among them, E2 represents the second trend change parameter of the target dimension in the target time sub-segment; D m represents the degree of weather resistance of the mth reference dimension of the target dimension; ρ m represents the Pearson correlation coefficient between the second vitality data sequence of the target dimension in the target time subsegment and the third vitality data sequence of the mth reference dimension in the target time subsegment; D m ×|ρ m| represents the data change correlation between the target dimension and the mth reference dimension in the target time subsegment; M represents the number of reference dimensions.

[0121] The first trend change parameter and the second trend change parameter can then be combined to obtain the comprehensive trend change parameter of the target dimension in the target time sub-segment. The same method as above can be used to obtain the comprehensive trend change parameter of the target dimension in each time sub-segment. The larger the comprehensive trend change parameter, the more obvious the change characteristics of the vitality data of the target dimension in each time sub-segment. The accumulated value of the comprehensive trend change parameters of the target dimension in all time sub-segments can then be used as the trend change degree of the target dimension. The greater the trend change degree, the more obvious the change characteristics of the vitality data of the target dimension within the preset time period.

[0122] In an embodiment of the present invention, the sum or product of the first trend change parameter and the second trend change parameter can be used as the comprehensive trend change parameter of the target dimension in the target time sub-segment to achieve the integration of the two, which is not limited here.

[0123] As an example, in one embodiment of the present invention, the expression of the trend change degree of the target dimension can be specifically, for example:

[0124]

[0125] Among them, E represents the degree of trend change of the target dimension; E (t,1) Indicates the first trend change parameter of the target dimension in the tth time subsegment; E (t,2) Indicates the second trend change parameter of the target dimension in the tth time sub-segment; E (t,1) +E (t,2) Represents the comprehensive trend change parameter of the target dimension in the tth time sub-segment; T represents the number of time sub-segments about the target dimension within the preset time period.

[0126] Due to the difference between the vitality data of the target dimension and the fitted vitality data every day in the preset time period, the interference factors in the overall trend of fitness exercises can be reflected, such as holiday arrangements or performances occupying stadiums. These interference factors often cannot truly reflect the manifestation of the vitality data of the target dimension on urban vitality. Therefore, it is necessary to analyze the difference between the vitality data of the target dimension and the fitted vitality data every day in the preset time period to obtain the interference possibility of the target dimension. The smaller the interference possibility, the less likely the fitness exercise reflected by the target dimension is to be interfered with, and thus the target dimension is more important in evaluating urban vitality.

[0127] Preferably, in one embodiment of the present invention, the method for obtaining the interference possibility of the target dimension specifically includes:

[0128] The absolute value of the difference between the vitality data of the target dimension and the fitted vitality data for each day in the preset time period is used as the fitting error of the target dimension for each day. The smaller the fitting error, the less likely the vitality data of the target dimension is to be interfered with. The average value of the fitting errors of the target dimension for all days in the preset time period can be used as the interference possibility of the target dimension.

[0129] As an example, in one embodiment of the present invention, the expression of the interference possibility of the target dimension can be specifically, for example, as follows:

[0130]

[0131] Where F represents the interference possibility of the target dimension; a r represents the vitality data of the target dimension on the rth day of the preset time period; a ′ r represents the fitted vitality data of the target dimension on the rth day of the preset time period; |a r -a ′ r | represents the fitting error of the target dimension in the rth day; R represents the number of days in the preset time period.

[0132] Finally, the greater the trend change of the target dimension and the smaller the interference possibility of the target dimension, the more important the target dimension is in the urban vitality evaluation. Therefore, the trend change degree can be used as the numerator and the interference possibility as the denominator. The comparison value can be normalized and the calculation result can be limited to the range of [0,1] to obtain the vitality evaluation weight of the target dimension.

[0133] As an example, in one embodiment of the present invention, the expression of the vitality evaluation weight of the target dimension can be specifically, for example:

[0134]

[0135] Among them, U represents the vitality evaluation weight of the target dimension; E represents the degree of trend change of the target dimension; F represents the interference possibility of the target dimension. Since there must be an error between the fitted data and the original data in the process of fitting the daily vitality data of the target dimension, F≠0; norm() represents the normalization function.

[0136] The vitality evaluation weight of each dimension can be obtained by the same method as above. Since the existing dimensionality reduction method only performs dimensionality reduction processing based on the statistical distribution characteristics of the data, it does not consider the importance of data of different dimensions to the evaluation of urban vitality, resulting in the loss of important dimensional data and the retention of unimportant dimensional data. The vitality evaluation weight can reflect the importance of each dimension in the evaluation of urban vitality. Therefore, according to the vitality evaluation weight of each dimension, the vitality data of all dimensions can be reduced in dimension to obtain the reduced dimensionality data, thereby improving the effect of data dimensionality reduction. Subsequently, a more effective evaluation of urban vitality can be carried out based on the reduced dimensionality data.

[0137] Preferably, in one embodiment of the present invention, the method for obtaining the data after dimensionality reduction specifically includes:

[0138] The vitality data and vitality evaluation weights of each dimension are input into the weighted PCA algorithm, and the weighted PCA algorithm is used to perform dimensionality reduction processing to obtain the reduced dimensionality data. The weighted PCA algorithm is a technical means well known to those skilled in the art and will not be described in detail here.

[0139] At this point, the dimensionality reduction processing of multi-dimensional vitality data is completed.

[0140] The city vitality evaluation module 104 is used to evaluate the vitality of the city to be tested based on the data after dimensionality reduction.

[0141] After the above-mentioned data dimensionality reduction processing, the obtained reduced-dimensional data is more important in the urban vitality evaluation. Therefore, the vitality evaluation of the tested city can be carried out based on the reduced-dimensional data, thereby improving the final effect of the vitality evaluation of the tested city.

[0142] Preferably, in one embodiment of the present invention, the method for evaluating the vitality of a city to be tested specifically includes:

[0143] The city vitality data set is searched on the Internet, and each city vitality data is scored and labeled by manual scoring, where the scoring values ​​are 0.1, 0.2, 0.3, ... 0.9, 1, a total of 10 score types. Then, the labeled city vitality data is trained with a neural network, and the reduced-dimensional data is input into the trained neural network to obtain the vitality score of the city to be tested, where the vitality score is the above-mentioned 10 scores. In one embodiment of the present invention, the neural network structure can select a five-layer fully connected neural network, and the loss function used in training is the cross-entropy function. The training of neural networks is a technical means well known to those skilled in the art and will not be elaborated here.

[0144] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0145] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. An urban vitality evaluation system based on multi-source dynamic big data fusion analysis, characterized by: The system comprises: The data collection module is used to obtain the vitality data of different dimensions of the city under test every day within a preset time period, and also obtain the meteorological data of different meteorological types every day; The weather impact analysis module is used to obtain daily weather indicators based on meteorological data of different weather types each day; obtain daily exercise suitability based on daily weather indicators and the difference between weather indicators of each day and adjacent days; and analyze the correlation between the vitality data of the target dimension and the exercise suitability of each day within a preset time period, taking any dimension as the target dimension, to obtain the target dimension's resistance to weather interference; The data dimensionality reduction module is used to obtain the vitality evaluation weight of the target dimension based on the changes in the vitality data of the target dimension on each day and the degree of resistance to weather interference of the target dimension, the correlation between the vitality data of the target dimension and other dimensions other than the target dimension on each day, and the degree of resistance to weather interference of other dimensions other than the target dimension; based on the vitality evaluation weight of each dimension, the vitality data of all dimensions are reduced in dimension to obtain the reduced data; The city vitality evaluation module is used to evaluate the vitality of the tested city based on the dimensionality-reduced data; Obtaining the degree of weather resistance of the target dimension includes: Fitting the exercise suitability of each day in a preset time period to obtain multiple time segments of the preset time period; Take any time segment as the target time segment, and sort the target dimension's vitality data for each day of the target time segment in chronological order as the first vitality data sequence for the target dimension in the target time segment; and take the target dimension's exercise fitness sequence for each day of the target time segment in chronological order as the exercise fitness sequence for the target time segment. Perform negative correlation mapping on the absolute value of the Pearson correlation coefficient between the first vitality data sequence and the exercise suitability sequence to obtain the weather interference resistance parameter of the target dimension in the target time segment; The two time segments closest to the target time segment are used as reference time segments. The difference between the target time segment and each reference time segment is used as the relative length difference between the target time segment and each reference time segment. The average of the relative length differences between the target time segment and all reference time segments is used as the numerator, the length of the target time segment is used as the denominator, and the value after normalization of the ratio is used as the weight parameter of the target dimension in the target time segment. Based on the weight parameter of the target dimension in each time segment, the weather resistance parameter of the target dimension in each time segment is weighted and summed to obtain the weather resistance degree of the target dimension; Obtaining the vitality evaluation weights of the target dimensions includes: Performing fitting processing on the vitality data of the target dimension on each day of a preset time period, and dividing the preset time period to obtain multiple time sub-segments of the preset time period regarding the target dimension, as well as fitting vitality data of the target dimension on each day of the preset time period; Take any time subsegment as the target time subsegment, take the absolute value of the difference in the target dimension's vitality data between each day in the target time subsegment and the next adjacent day as the target dimension's vitality data change for each day in the target time subsegment, and take the product of the average value of the target dimension's vitality data change for all days in the target time subsegment and the target dimension's weather interference resistance as the first trend change parameter of the target dimension in the target time subsegment; Taking all dimensions other than the target dimension as reference dimensions, the correlation between the target dimension and the vitality data of each reference dimension on each day in the target time segment is analyzed. Combined with the degree of weather resistance of each reference dimension, the second trend change parameter of the target dimension in the target time segment is obtained. The first trend change parameter and the second trend change parameter are combined to obtain the comprehensive trend change parameter of the target dimension in the target time sub-segment, and the cumulative value of the comprehensive trend change parameter of the target dimension in all time sub-segments is used as the trend change degree of the target dimension; Obtaining the interference possibility of the target dimension based on the difference between the vitality data of the target dimension and the fitted vitality data in each day of the preset time period; Take the degree of trend change as the numerator and the possibility of interference as the denominator, normalize the comparison value, and obtain the vitality evaluation weight of the target dimension.

2. The urban vitality evaluation system based on multi-source dynamic big data fusion analysis according to claim 1 is characterized in that: The method of obtaining daily weather indicators includes: Take any day as the target day, and take the absolute value of the difference between the meteorological data of each meteorological type on the target day and the standard suitable data of each meteorological type as the meteorological deviation degree of each meteorological type on the target day; The weather deviation degrees of all weather types on the target day are integrated and normalized to obtain the weather index of the target day.

3. The urban vitality evaluation system based on multi-source dynamic big data fusion analysis according to claim 2 is characterized in that: Obtaining daily exercise fitness includes: The absolute value of the difference between the weather index on the target day and the previous day is used as the weather index change on the target day; The weather index and the change amount of the weather index on the target day are integrated and negatively correlated to obtain the exercise suitability on the target day.

4. The urban vitality evaluation system based on multi-source dynamic big data fusion analysis according to claim 1 is characterized in that: The step of fitting the exercise suitability of each day in a preset time period to obtain a plurality of time segments of the preset time period includes: Using the least square method, curve fitting is performed on the two-dimensional data points constituted by the daily exercise fitness to obtain a fitting curve; The extreme points on the fitting curve are used as segmentation points, and the preset time period is divided into multiple time segments using each segmentation point.

5. The urban vitality evaluation system based on multi-source dynamic big data fusion analysis according to claim 1 is characterized in that: The step of obtaining the second trend change parameter of the target dimension in the target time subsegment includes: Take any reference dimension as the target reference dimension, sort the vitality data of the target dimension in each day of the target time subsegment in time sequence as the second vitality data sequence of the target dimension in the target time subsegment, and take the vitality data of the target reference dimension in each day of the target time subsegment in time sequence as the third vitality data sequence of the target reference dimension in the target time subsegment; The product of the absolute value of the Pearson correlation coefficient between the second vitality data sequence and the third vitality data sequence and the weather interference resistance level of the target reference dimension is used as the data change correlation between the target dimension and the target reference dimension in the target time subsegment; The accumulated value of the data change correlation between the target dimension and all reference dimensions in the target time subsegment is used as the second trend change parameter of the target dimension in the target time subsegment.

6. The urban vitality evaluation system based on multi-source dynamic big data fusion analysis according to claim 1 is characterized in that: The interference possibilities for obtaining the target dimension include: The absolute value of the difference between the vitality data and the fitted vitality data of the target dimension in each day of the preset time period is used as the fitting error of the target dimension in each day; The average value of the fitting errors of the target dimension in all days within a preset time period is used as the interference possibility of the target dimension.

7. The urban vitality evaluation system based on multi-source dynamic big data fusion analysis according to claim 1 is characterized in that: The data obtained after dimensionality reduction includes: The vitality data of each dimension and the vitality evaluation weight are input into a weighted PCA algorithm, and the weighted PCA algorithm is used to perform dimensionality reduction processing to obtain data after dimensionality reduction.

8. The urban vitality evaluation system based on multi-source dynamic big data fusion analysis according to claim 1 is characterized in that: The vitality evaluation of the tested city includes: The urban vitality data set was searched on the Internet, and each urban vitality data was scored and labeled by manual scoring. A neural network was trained on the labeled urban vitality data, and the dimensionality-reduced data was input into the trained neural network to obtain the vitality score of the tested city.

Images