Dynamic control value control method and device based on data fitting, equipment and medium

Abstract

The invention relates to the technical field of intelligent decision making, can be applied to business system platforms of financial science and technology, medical health and the like, and discloses a dynamic control value control method and device based on data fitting, equipment and a medium, and the method comprises the steps: obtaining a feature factor of a preset type of service of a user, and constructing a spatial-temporal feature tensor according to the feature factor; calculating a dynamic characteristic range according to the spatial-temporal characteristic tensor; constructing a behavior characteristic matrix through user historical behavior data obtained in advance, and determining a characteristic evaluation index according to the dynamic characteristic range and the behavior characteristic matrix; and performing additional control value fitting on the feature evaluation index according to pre-acquired historical reference data to obtain an additional control value. By analyzing the characteristic factors of the preset type service of the user, the calculated adjustment control value is more accurate, and large deviation of a calculation result caused by only considering a single element or static information is avoided.

Description

Technical Field

[0001] The present invention relates to the field of intelligent decision-making technology, and in particular to a dynamic control value control method, device, equipment and medium based on data fitting. Background Art

[0002] At present, with the continuous expansion of mobile Internet application scenarios and the iteration of software technology, many financial technology application products and medical and health application products are constantly entering people's work and life.

[0003] For example, in the fintech sector, there are numerous financial products, financing products, and travel insurance products targeting individual users. These products not only allow people to conduct business online but also provide differentiated financial product services tailored to the specific circumstances of different individuals. Another example is healthcare, where there are numerous accident insurance and disease insurance products targeting individual users. These products also not only allow people to conduct business online but also provide differentiated healthcare products and services tailored to the specific circumstances of different individuals.

[0004] Currently, the fintech and healthcare applications currently available on the market provide differentiated services tailored to the specific needs of different individuals, often based on specific dynamic control values. (For example, different individuals may have different tolerances for the characteristics of financial products, and the characteristic tolerance data calculated for each individual through technical means can serve as the corresponding dynamic control value for differentiated services.) However, the characteristic control calculations currently performed for individuals are typically based on rule-based models combined with human experience. The calculation process is often based on individual historical data and is easily influenced by subjective factors, making it difficult to guarantee the accuracy and objectivity of the results. Summary of the Invention

[0005] The present invention provides a dynamic control value control method, device, equipment and medium based on data fitting. By analyzing the characteristic factors of the user's preset type of service, the calculated adjustment control value is made more accurate, avoiding large deviations in the calculation results caused by only considering a single element or static information.

[0006] In a first aspect, a dynamic control value control method based on data fitting is provided, comprising:

[0007] Obtaining characteristic factors of a user-preset type of service, and constructing a spatiotemporal feature tensor based on the characteristic factors;

[0008] Calculating a dynamic feature range based on the spatiotemporal feature tensor;

[0009] Constructing a behavior feature matrix by using pre-acquired user historical behavior data, and determining a feature evaluation index according to the dynamic feature range and the behavior feature matrix;

[0010] Performing additional control value fitting on the characteristic evaluation index according to pre-acquired historical reference data to obtain an additional control value;

[0011] Generate a dynamic control value based on the combination of the additional control value and the preset basic control value;

[0012] A user preset type message is obtained in real time or periodically, and the dynamic control value is numerically adjusted according to the obtained user preset type information to obtain an adjusted control value.

[0013] In a second aspect, a dynamic control value control device based on data fitting is provided, comprising:

[0014] An acquisition module is used to obtain characteristic factors of a user-preset type of service;

[0015] A construction module, configured to construct a spatiotemporal feature tensor according to the feature factors;

[0016] A calculation module, configured to calculate a dynamic feature range based on the spatiotemporal feature tensor;

[0017] A matrix construction module is used to construct a behavior feature matrix based on pre-acquired user historical behavior data;

[0018] A determination module, configured to determine a feature evaluation index based on the dynamic feature range and the behavior feature matrix;

[0019] A fitting module, configured to perform additional control value fitting on the characteristic evaluation index according to pre-acquired historical reference data to obtain an additional control value;

[0020] A generating module, configured to generate a dynamic control value based on the additional control value and a preset basic control value;

[0021] The acquisition and adjustment module is used to acquire user preset type information in real time or periodically, and to adjust the dynamic control value according to the acquired user preset type information to obtain an adjusted control value.

[0022] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned dynamic control value control method based on data fitting are implemented.

[0023] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned dynamic control value control method based on data fitting are implemented.

[0024] In the above-mentioned scheme implemented by the dynamic control value control method, device, computer equipment and storage medium based on data fitting, the characteristic factors of the user's preset type service are analyzed to make the calculated adjustment control value more accurate, avoiding large deviations in the calculation results due to only considering a single element or static information. The system analyzes the relevant information of the preset area in each user's preset type service, for example, travel conditions, transportation methods, weather conditions and other multi-dimensional dynamic factors to perform risk assessment and adjust the control value calculation to achieve truly personalized pricing. At the same time, through the real-time data update mechanism, the system re-evaluates the user's risk and adjusts the control value based on the obtained user preset type information, for example, based on the latest weather, public opinion and other information, to ensure the rationality and accuracy of the calculation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. 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 labor.

[0026] Figure 1 This is a schematic diagram of an application environment of a dynamic control value control method based on data fitting in one embodiment of the present invention;

[0027] Figure 2 This is a flow chart of a dynamic control value control method based on data fitting in one embodiment of the present invention;

[0028] Figure 3 1 is a structural diagram of a dynamic control value control device based on data fitting in one embodiment of the present invention;

[0029] Figure 4 is a structural diagram of a computer device according to an embodiment of the present invention;

[0030] Figure 5 It is another structural schematic diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0031] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0032] The embodiment of the present invention provides a dynamic control value control method based on data fitting, which can be applied in the following situations: Figure 1 In an application environment, the client communicates with the server via a network. The server can obtain characteristic factors of user-preset type services, construct a spatiotemporal feature tensor based on the characteristic factors, calculate a dynamic feature range based on the spatiotemporal feature tensor, construct a behavior feature matrix based on pre-acquired user historical behavior data, and determine a feature evaluation index based on the dynamic feature range and the behavior feature matrix; perform additional control value fitting on the feature evaluation index based on pre-acquired historical reference data to obtain an additional control value; generate a dynamic control value based on the additional control value and a preset basic control value combination; obtain user-preset type messages in real time or at a fixed time, numerically adjust the dynamic control value based on the obtained user-preset type information to obtain an adjusted control value, and feed the adjusted control value back to the client. The present invention provides a dynamic control value control device based on data fitting. For the adjustment control value service, the device analyzes the characteristic factors of the user-preset type service to make the calculated adjustment control value more accurate, avoiding large deviations in the calculation results caused by only considering a single element or static information. The client can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices. The server can be implemented as an independent server or a server cluster consisting of multiple servers. The present invention is described in detail below through specific examples.

[0033] See also Figure 2 As shown, Figure 2 A flow chart of a dynamic control value control method based on data fitting provided by an embodiment of the present invention includes the following steps:

[0034] S1. Obtain characteristic factors of a user-preset type of service, and construct a spatiotemporal feature tensor based on the characteristic factors.

[0035] In an embodiment of the present invention, the acquisition refers to identifying key factors that may affect the probability and amount of compensation through data collection, analysis and evaluation, and the construction refers to integrating spatiotemporal data according to given characteristic factors to create a tensor representing spatiotemporal risk.

[0036] Specifically, user-preset type services include travel accident insurance, financial insurance, life service insurance and other services. Characteristic factors refer to risk factors. Taking travel accident insurance as an example, characteristic factors include weather data, public opinion, etc. Through real-time analysis of users' travel data (such as transportation, travel date, destination, weather, public opinion, etc.), characteristic factors are organized and integrated in time and space dimensions to form a multi-dimensional data structure, so as to better analyze and understand the risk situations at different times and spaces.

[0037] In specific medical and health scenarios, based on the characteristic factors of travel accident insurance, such as travel routes, personnel flow patterns, etc., combined with historical disease data, the trend of disease outbreaks that may occur in specific areas within a specific time period can be predicted, so that preventive measures can be taken in advance. For example, before the peak tourist season arrives, medical institutions near popular tourist attractions should stockpile corresponding medicines and medical equipment, and strengthen health promotion and vaccination for tourists.

[0038] In the FinTech scenario, by constructing a spatiotemporal feature tensor, these factors can be quantitatively analyzed in the time and space dimensions. For example, for tourists traveling to high-risk areas (such as war-torn areas or areas prone to natural disasters), during a specific time period (such as the peak tourist season), based on the high-risk situation obtained by the spatiotemporal feature tensor analysis, insurance companies can reasonably increase insurance premiums, achieve differentiated pricing, and ensure that the pricing of insurance products matches the actual risks.

[0039] In the embodiment of the present invention, the step of obtaining characteristic factors of a user-preset service type includes:

[0040] Determine a user preset area according to the user preset type service, and obtain environmental feature data according to the user preset area at a set time;

[0041] Using a preset traffic system to obtain traffic condition data according to the user preset area;

[0042] Using public opinion information to obtain security dynamic data based on the user's preset area;

[0043] The environmental characteristic data, traffic condition data and safety dynamic data are aggregated into characteristic factors.

[0044] In the embodiment of the present invention, the acquisition of environmental characteristic data refers to calling a meteorological API (application programming interface) to obtain meteorological data of the travel departure and destination, and the acquisition of traffic condition data and security dynamic data is the same as the above steps.

[0045] Specifically, the user preset area is determined according to the user preset type service. In this solution, it means determining the travel location according to the user preset type service. Then the system collects dynamic data that affects the user's travel safety from multiple external data sources. Among them, environmental characteristic data includes weather risk data, terrain risk data, etc. For example, taking weather risk data as an example, the meteorological data of the travel departure and destination are obtained through the meteorological API, the traffic information system obtains traffic conditions and congestion, and the public opinion data source collects the safety dynamics of the destination, such as whether there have been recent accidents, social unrest, etc.; the characteristic factors mainly include transportation methods (such as airplanes, trains, cars), weather conditions (such as sunny days, rainy days, blizzards, etc.), destination safety (such as whether there have been terrorist attacks, riots and other social public opinions in the near future), and the crowd density of the destination (such as holidays and popular destinations are bound to have high crowd density) are updated in real time through the data collection module.

[0046] In an embodiment of the present invention, constructing a spatiotemporal feature tensor based on the feature factors includes:

[0047] Quantitatively processing the environmental characteristic data, traffic condition data, and safety dynamic data to obtain quantitative data;

[0048] Extracting the time dimension and space dimension of the characteristic factor, and determining the spatiotemporal dimension according to the time dimension and space dimension;

[0049] Constructing an initial spatiotemporal feature tensor according to the spatiotemporal dimension and the number of the feature factors;

[0050] The quantized data is filled into the initial spatiotemporal feature tensor to obtain the spatiotemporal feature tensor.

[0051] In an embodiment of the present invention, the quantization processing refers to the process of converting environmental feature data, traffic condition data and safety dynamic data that are originally non-numerical or not convenient for direct mathematical operations and analysis into numerical form or numerical analysis through certain methods. The extraction refers to separating information related to time and space from the characteristic factor data for further analysis and modeling. The determination refers to determining the dimension that describes the spatiotemporal characteristics of the data set, that is, the spatiotemporal dimension. The construction of the initial spatiotemporal feature tensor refers to using the spatiotemporal dimension and the number of characteristic factors to create a multidimensional data structure (tensor) for storing and representing the distribution and changes of risks in time and space. The filling refers to filling the quantitative data or information into the corresponding position or dimension of the initial spatiotemporal feature tensor to improve or complete this multidimensional data structure.

[0052] Specifically, the temperature data included in the environmental characteristic data can be normalized by mapping the actual temperature value to a range of 0-1. Assuming that the highest temperature in history is T max , the lowest temperature is Tmin , the current temperature is T, then the quantized temperature value T q =(TT min ) / (T max -T min ), for precipitation, it is graded and quantified according to the range of precipitation, such as setting 0-5 mm as level 1, 5-15 mm as level 2, 15-30 mm as level 3, and above 30 mm as level 4; for wind speed, it can also be graded and quantified according to the degree of its impact on tourism activities; for example, 0-3 m / s is level 1, 3-6 m / s is level 2, 6-10 m / s is level 3, and above 10 m / s is level 4; the degree of road congestion can be quantified by calculating the utilization rate of road capacity. Assuming that the design capacity of a certain section of road is C and the actual traffic volume is Q, the congestion index I = Q / C, and the value of I is mapped to the interval of 0-1, 0 means smooth traffic and 1 means severe congestion. The traffic accident rate can be quantified based on the number of traffic accidents in a certain area within a certain period of time (such as one month). If the area has a maximum of N accidents per month in history, max If there are N accidents in the current month, then the quantified accident rate R = N / N max .

[0053] Furthermore, safety incidents at tourist attractions can be graded and quantified based on the nature and impact of the incidents. For example, a minor tourist fall is recorded as Level 1, a tourist injury due to facility failure is recorded as Level 2, and a major safety accident involving multiple people is recorded as Level 3. The social security situation can be assessed by conducting a questionnaire survey on tourists, asking them to rate the local security situation on a scale of 1-5, with 1 representing very poor and 5 representing very good. The scoring results are then statistically averaged to obtain a quantitative value of social security in the area.

[0054] Specifically, the time dimension can be determined based on the frequency of data collection and analysis requirements. For example, if data is collected daily, each day can be considered a time unit and represented by an integer, such as Day 1, Day 2, etc. If seasonal risks are being analyzed, each season can also be considered a time unit, with 1-4 representing spring, summer, autumn, and winter, respectively. The spatial dimension needs to be divided according to the geographic scope of the data. A tourist area can be divided into several small geographic units, such as a tourist attraction as the center and the surrounding area divided into circular areas of different radii, or the entire tourist city can be divided into several blocks. Each geographic unit is represented by a unique number, such as S1, S2, etc. The spatiotemporal dimension combines the time dimension and the spatial dimension to form a two-dimensional space-time coordinate. For example, the Sith area on the tth day can be represented as (t, Si).

[0055] Furthermore, assuming that n characteristic factors are determined through analysis, the characteristic factors include risk factors such as temperature, precipitation, wind speed, road congestion, traffic accident rate, scenic spot safety incident level, social security score, etc. The initial spatiotemporal feature tensor is a three-dimensional tensor. The spatiotemporal feature tensor includes a spatiotemporal risk tensor with dimensions of time, space, and characteristic factors. It can be represented by a three-dimensional array, such as R[t,Si,j], where t represents the time dimension, Si represents the space dimension, and j represents the index of the characteristic factor, j=1,2,…,n. In the initial spatiotemporal feature tensor, each element is initialized to a default value, such as 0 or a null value, indicating that specific quantitative data has not yet been filled in. The quantized data is filled into the initial spatiotemporal feature tensor according to its corresponding spatiotemporal dimension and characteristic factor index. For example, the temperature quantization value of the Sith region on the tth day is T q , it is the first characteristic factor, then fill Tq into the R[t,Si,1] position of the tensor; similarly, fill the quantized precipitation value of the area on that day into R[t,Si,2], and the quantized wind speed value into R[t,Si,3], and so on, until all the quantized data are filled into the corresponding positions, and finally obtain a complete spatiotemporal feature tensor.

[0056] In the embodiment of the present invention, by clarifying the time dimension and the space dimension, the distribution of risks at different time points and different geographical locations can be accurately pointed out. At the same time, in the form of tensors, the comprehensive information of characteristic factors in different time and space dimensions can be comprehensively reflected.

[0057] In the embodiment of the present invention, the spatiotemporal feature tensor can integrate these different types of feature factors together, taking into account both the time and space dimensions, and provide a comprehensive and three-dimensional description of the risks in the travel process.

[0058] S2. Calculate the dynamic feature range according to the spatiotemporal feature tensor.

[0059] In the embodiment of the present invention, the calculation refers to the process of using specific mathematical methods and models to process and analyze the information contained in the spatiotemporal feature tensor to obtain the dynamic change range of the risk in time and space.

[0060] Specifically, the dynamic feature range includes the dynamic risk range. Data related to each feature factor is extracted from the spatiotemporal feature tensor, and its changing patterns in time and space are analyzed. Common models including probability models, machine learning models, geographic information system (GIS) models, etc. are used to analyze the distribution characteristics and diffusion trends of risks in geographic space, and the calculated dynamic feature range is presented in an intuitive way, such as drawing risk maps, time series graphs, risk matrices, etc.

[0061] In healthcare scenarios, analyzing characteristic factors such as disease incidence data across different regions and time periods contained in spatiotemporal feature tensors can promptly detect unusual disease transmission patterns and potential epidemic trends. For example, during an epidemic, monitoring case data from different cities and communities, combined with time series analysis, can identify rising trends in the number of cases and provide early warning for disease prevention and control.

[0062] At the same time, the spatiotemporal feature tensor can integrate information from multiple dimensions, including the borrower's geographic location, local economic development, industry distribution, and time-series economic data fluctuations. By analyzing this information, a more comprehensive and detailed credit risk profile can be constructed for the borrower.

[0063] In the embodiment of the present invention, the step of calculating the dynamic feature range based on the spatiotemporal feature tensor includes:

[0064] Performing aggregation analysis on the spatiotemporal feature tensor to obtain aggregated spatiotemporal feature data;

[0065] Calculating the standard deviation of characteristic factors one by one according to the aggregated spatiotemporal characteristic data, and measuring the degree of dispersion of the characteristic factors by the standard deviation;

[0066] A dynamic feature range coefficient of the feature factor is set, and a dynamic feature range is determined according to the discrete degree and the dynamic feature range coefficient.

[0067] In an embodiment of the present invention, the aggregation analysis refers to summarizing, integrating or calculating the data in the spatiotemporal feature tensor according to a certain rule or method to generate a higher-level summary result. The calculation refers to calculating the standard deviation of the data of each characteristic factor in the aggregated spatiotemporal feature data through mathematical methods or computing techniques. The setting refers to setting a range coefficient for each characteristic factor according to a specific risk management strategy or standard, which is used to measure the risk level or variation range of the characteristic factor within a specified range.

[0068] Specifically, the spatiotemporal feature tensor quantified in step S1 is aggregated to integrate the risk information in the spatiotemporal dimension. For example, if the original data is recorded in days and we want to obtain the monthly risk situation, then the daily risk data in each month can be summed, averaged, or other statistical operations can be performed. For the spatial dimension, aggregation can be performed according to the division of geographical areas or grids; for example, several adjacent grid cells can be merged into a larger area, and the risk data in the area can be summed, averaged, or weighted averaged. Based on the aggregated spatiotemporal feature data, the standard deviation or coefficient of variation of the characteristic value of each spatiotemporal point is calculated to measure the degree of discreteness of the risk and thus determine the dynamic feature range.

[0069] In detail, in time aggregation, assuming that we adopt the averaging method and perform time aggregation in months, then the aggregated eigenvalue R′ of the mth month is m,s,f , expressed by the following formula:

[0070]

[0071] Among them, D m is the number of days in the mth month, and t∈m means that time point t belongs to the mth month.

[0072] Furthermore, suppose that area A contains n grid cells, and the weight of each grid cell s is W s , then the aggregated eigenvalue R′ of region A t,A,f , expressed by the following formula:

[0073] R′ t,A,f =∑ s∈A w s R′ t,s,f

[0074] Among them, ∑ s∈A w s =1.

[0075] Furthermore, suppose that at a certain time point t and in region A, the aggregated eigenvalue of the characteristic factor f is R′ t,A,f , whose mean is The risk standard deviation at this time and space point is Where N is the number of samples used to calculate the mean. By setting a certain multiple (such as K = 2 or k = 3), the dynamic feature range can be obtained as

[0076] In an embodiment of the present invention, through aggregation analysis, data can be observed at a higher level to discover potential patterns, trends and relationships that are difficult to detect in a single data point. By calculating the standard deviation of the characteristic factor, the fluctuation of each characteristic factor can be accurately quantified, and the dynamic characteristic range can be determined by combining the discrete degree of the characteristic factor and the dynamic characteristic range coefficient, which can comprehensively consider the volatility of the data and the business's risk tolerance.

[0077] In an embodiment of the present invention, by integrating the information in the spatiotemporal feature tensor to calculate the dynamic feature range, the true situation of the risk can be reflected more comprehensively and accurately, avoiding risk assessment deviations caused by considering only a single factor or static information.

[0078] S3. Construct a behavior feature matrix using pre-acquired user historical behavior data, and determine a feature evaluation index based on the dynamic feature range and the behavior feature matrix.

[0079] In an embodiment of the present invention, the construction refers to presenting the user's risk behavior in the form of a matrix through pre-acquired user historical behavior data, and the determination refers to evaluating and classifying the risk level of the user or event based on the data and information provided by the dynamic feature range and the behavior feature matrix, and finally determining its feature evaluation index.

[0080] Specifically, user historical behavior data includes user historical risk records. The dimensions of the behavior feature matrix are determined based on the analysis needs and relevant factors. The behavior feature matrix includes a risk behavior matrix. For each user, specific values ​​are filled in the corresponding matrix positions based on their historical risk records. Each element in the behavior feature matrix is ​​given a certain weight, which is determined based on the degree of its impact on the overall risk. At the same time, the feature evaluation index represents the risk assessment level.

[0081] In specific healthcare scenarios, medical institutions can rationally allocate medical resources based on patient characteristic assessment indicators. Patients with high characteristic assessment indicators are prioritized for resources such as expert consultations, hospital beds, and advanced testing equipment. By analyzing the distribution of characteristic assessment indicators among patients within a region, health management departments can optimize the layout of medical services.

[0082] In financial scenarios, insurance companies can set differentiated insurance product prices based on characteristic evaluation indicators. For example, for auto insurance, if a policyholder has multiple traffic accident claims (historical accident record) and lives in an area with congested traffic and a high accident rate (dynamic characteristic range), and is assessed as having high characteristic evaluation indicators, then their auto insurance control value will be relatively high. On the other hand, policyholders with low characteristic evaluation indicators, such as a good driving record and good traffic conditions in their area, can enjoy a lower control value discount. This ensures fair and reasonable insurance pricing, ensuring that insurance companies can achieve profitability while keeping risks under control.

[0083] In the embodiment of the present invention, determining the feature evaluation index according to the dynamic feature range and the behavior feature matrix includes:

[0084] According to the behavior feature matrix, weights are assigned to the historical feature factors in the user's historical behavior data one by one to obtain historical weights corresponding to the historical feature factors;

[0085] According to the dynamic feature range, the feature factors are weighted one by one to obtain feature weights corresponding to the feature factors;

[0086] A comprehensive feature value is calculated based on the historical weight and the feature weight, and a feature evaluation index is determined based on the comprehensive feature value.

[0087] In an embodiment of the present invention, the weight allocation refers to analyzing and evaluating the historical characteristic factors in the user's historical behavior data one by one according to the behavior characteristic matrix, and assigning a weight value to each characteristic factor to reflect its contribution or importance to the user's risk behavior. The characteristic factor weight allocation is similar. The calculation refers to using the data of historical weights and characteristic weights to quantify the user's comprehensive characteristic value through certain rules or methods.

[0088] Specifically, the feature weight includes the risk weight, and the comprehensive feature value can also be expressed as a comprehensive risk value. For each user, a specific value is filled in the corresponding matrix position based on their historical accident records. For example, if a user has had three accidents in the past year, "3" is filled in the row corresponding to the "Number of Accidents" column. If the proportion of accidents caused by drunk driving is 20%, "0.2" is filled in the row corresponding to the "Proportion of Accidents Caused by Drunk Driving" column. In this way, a behavioral feature matrix is ​​constructed for each user, and the range of each feature factor at the current moment is determined based on the dynamic feature range calculated previously. For example, if the dynamic feature range of a feature factor is [30,70], and the current value of the factor is 60, it can be seen that it is in a higher-risk interval.

[0089] Furthermore, for each user, the standardized behavioral feature matrix elements are multiplied by their corresponding weights and summed to obtain the behavioral feature matrix score S1. At the same time, the dynamic feature range score S2 is calculated based on the current value of each feature factor in the dynamic feature range and its weight; assuming that the standardized value of the i-th element in the behavioral feature matrix is ​​x i , the value of the jth characteristic factor in the dynamic characteristic range is y j , then the formula is:

[0090]

[0091] Among them, w i refers to the historical weight of the i-th element, v j It refers to the feature weight of the j-th element.

[0092] Finally, S1 and S2 are weighted and summed in a certain ratio to obtain the comprehensive eigenvalue S, for example, S = αS1 + (1-α)S2, where α is the weight coefficient, which can be determined according to actual conditions and has a value range of [0,1].

[0093] In an embodiment of the present invention, by assigning weights to historical characteristic factors through a behavioral characteristic matrix, the degree of influence of different characteristic factors on the risk status of each user can be accurately reflected based on the unique risk history of each user. The weight assignment of characteristic factors can timely reflect the degree of influence of these macro factors on different characteristic factors.

[0094] In this embodiment of the present invention, the dynamic feature range can promptly reflect the impact of real-time changes in the macro environment on user risk, while the behavioral feature matrix is ​​continuously updated with new historical risk records. This dynamic combination captures the dynamic changes in a user's risk profile, enabling timely adjustment of feature assessment indicators to more accurately reflect the user's current true risk level.

[0095] S4. Perform additional control value fitting on the characteristic evaluation index according to pre-acquired historical reference data to obtain an additional control value.

[0096] In the embodiment of the present invention, the control value fitting refers to further adjusting and optimizing the control values ​​of the determined characteristic evaluation indicators based on historical reference data, so as to more accurately reflect the actual risk level and potential degree of loss of the insured object.

[0097] Specifically, historical reference data includes historical claims data, historical accident data, etc. After the characteristic evaluation index is evaluated, it is dynamically adjusted based on the characteristic evaluation index R. Linear regression can fit the relationship between the characteristic evaluation index R and the additional control value through historical reference data. Among them, the control value can refer to premiums, premium increase coefficients, premium decrease coefficients and other data, and the additional control value includes the risk additional control value.

[0098] In the specific context of healthcare, the setting of additional control values ​​can, to a certain extent, mitigate moral hazard. When insured persons understand that their risky behaviors or health conditions can affect their control values, they may pay more attention to their health management and reduce unnecessary medical consumption. For example, smokers may consider quitting because they have to pay a higher additional control value, thereby reducing their characteristic assessment indicators and control value expenditures.

[0099] In the fintech scenario, fintech companies can understand the probability of compensation and the distribution of compensation amounts for customers with different characteristic evaluation indicators by analyzing historical reference data. For high-risk customers, higher additional control values ​​are charged, while for low-risk customers, lower control values ​​are charged. This allows the control value to match the customer's actual risk level, ensuring that the company obtains reasonable returns while taking risks.

[0100] In the embodiment of the present invention, performing additional control value fitting on the feature evaluation index based on pre-acquired historical reference data to obtain the additional control value includes:

[0101] Perform linear regression analysis on the historical reference data obtained in advance to obtain the regression coefficient;

[0102] Combining the regression coefficient with the characteristic evaluation index to obtain a combined formula;

[0103] Assigning a value to the feature evaluation index to obtain a feature value;

[0104] The combined formula is solved for an additional control value according to the regression coefficient and the eigenvalue to obtain an additional control value.

[0105] In an embodiment of the present invention, the linear regression analysis refers to finding the relationship between independent variables (which may be characteristic factors, characteristics, etc.) and dependent variables (compensation amount, frequency, etc.) by establishing a linear model, and obtaining the regression coefficient of each independent variable. The formula combination refers to combining the regression coefficient obtained by the linear regression analysis with the characteristic evaluation index to construct a comprehensive mathematical formula to describe the relationship between the characteristic evaluation index and the compensation data. The assignment refers to assigning corresponding numerical values ​​to the characteristic evaluation index according to low, medium and high risks. The additional control value solution refers to the process of substituting the corresponding numerical values ​​and regression coefficients assigned to low, medium and high risks into the combination formula to obtain the final result.

[0106] Specifically, we collect historical reference data, which usually contains multiple variables, such as the amount of compensation paid, the number of compensation payments, the age of the insured, the insurance period, and the insurance amount. Suppose we want to study the relationship between the amount of compensation paid and the age of the insured, then the amount of compensation paid is the dependent variable y, and the age of the insured is the independent variable x; we classify the characteristic evaluation index R into low risk, medium risk, and high risk:

[0107] When the characteristic evaluation index is low risk, R = 1;

[0108] When the characteristic evaluation index is medium risk, R = 2;

[0109] When the feature evaluation index is high risk, R=3.

[0110] Furthermore, the general form of the linear regression model is y = α0 + α1x + ε, where α0 refers to the intercept, α1 refers to the regression coefficient, and ε refers to the error term. The regression coefficient is solved by the least squares method, and the relationship between the characteristic evaluation index R and the additional control value is fitted by historical reference data. The calculation formula is as follows:

[0111] Additional control value = α0 + α1R + α2R 2 +···α n R n

[0112] Among them, α0, α1, ···α n It refers to the regression coefficient, R refers to the corresponding value of the characteristic evaluation index, and the characteristic value represents the risk value.

[0113] In an embodiment of the present invention, the regression coefficient is obtained through linear regression analysis, which can accurately quantify the linear relationship between each factor in the historical reference data and the compensation result. The analysis based on the actual historical data makes the decision more scientific and objective, and reduces the influence of human subjective judgment. Since the regression coefficient is dynamically calculated based on historical data, when new data is added, the regression coefficient will change accordingly, thereby enabling the combined formula to timely reflect the changing trend of risk factors.

[0114] In this embodiment of the present invention, based on analysis of historical reference data, it is possible to clearly identify which customers belong to the high-risk group. By charging a higher additional control value, this can, to a certain extent, discourage high-risk customers from purchasing insurance, or encourage them to take measures to reduce their risk after being underwritten, thereby reducing potential claims losses.

[0115] S5. Generate a dynamic control value based on the combination of the additional control value and the preset basic control value.

[0116] In the embodiment of the present invention, the combination generation refers to a process of obtaining a result by adding the additional control value and the basic control value.

[0117] Specifically, the dynamic control value represents the risk control value. After the characteristic evaluation index is evaluated, the system will dynamically adjust the control value based on the user's characteristic evaluation index R. The regression model is mainly used here to calculate the final control value amount, where the basic control value refers to the standardized pricing for different insurance products.

[0118] Linear regression model: Based on the output of the characteristic evaluation index, the system can use linear regression to calculate the control value. The calculation formula of the dynamic control value is as follows:

[0119] Dynamic control value = basic control value + additional control value.

[0120] S6. Obtain user preset type information in real time or periodically, and adjust the dynamic control value according to the obtained user preset type information to obtain an adjusted control value.

[0121] In the embodiment of the present invention, the numerical adjustment refers to the insurance company re-evaluating and modifying the originally determined dynamic control value based on the upcoming travel information obtained in advance to obtain an adjusted control value that is more in line with the actual risk situation.

[0122] Specifically, user preset type messages are obtained in real time or at regular intervals. User preset type messages include temporary travel weather data, temporarily changed transportation methods, or temporarily changed public opinion messages, etc. The system collects characteristic factors in real time. For example, some time before the user travels, the system regularly checks the latest weather, road conditions, public opinion and other information, re-evaluates the user's characteristic evaluation indicators, and adjusts the control value according to the new risk assessment results. If the user's risk increases, the control value is increased; if the risk decreases, the control value is reduced.

[0123] In the embodiment of the present invention, the step of numerically adjusting the dynamic control value according to the acquired user preset type information to obtain the adjusted control value includes:

[0124] Re-evaluate the feature evaluation index according to the obtained user preset type information to obtain an evaluation result;

[0125] The dynamic control value is numerically adjusted according to the evaluation result to obtain an adjusted control value.

[0126] In an embodiment of the present invention, the assessment refers to the insurance company re-analyzing and judging the characteristic assessment indicators of the insured based on the user preset type information obtained, using specific methods and models, to determine the degree of risk faced by the insured in the upcoming travel.

[0127] Specifically, based on the user preset type information obtained, the latest characteristic factors affecting the pricing of travel accident insurance are collected through meteorological API, traffic system and other equipment. The characteristic factors are analyzed through steps S2 and S3, and the characteristic evaluation indicators are calculated. The latest characteristic evaluation indicators are substituted into the combination formula to obtain the adjusted control value.

[0128] In an embodiment of the present invention, through a real-time data update mechanism, the system can reassess the user's risk and adjust the control value based on user preset type information, for example, based on the latest weather, public opinion and other information, to ensure the rationality and accuracy of pricing.

[0129] It can be seen that in the above scheme, for the adjustment control value service, the characteristic factors of the user preset type service are obtained, and a spatiotemporal characteristic tensor is constructed according to the characteristic factors; the dynamic characteristic range is calculated according to the spatiotemporal characteristic tensor; a behavior characteristic matrix is ​​constructed through the pre-acquired user historical behavior data, and a characteristic evaluation index is determined according to the dynamic characteristic range and the behavior characteristic matrix; an additional control value is fitted to the characteristic evaluation index according to the pre-acquired historical reference data to obtain an additional control value; a dynamic control value is generated according to the combination of the additional control value and the preset basic control value; the user preset type message is obtained in real time or at a fixed time, and the dynamic control value is numerically adjusted according to the obtained user preset type information to obtain an adjusted control value. By analyzing the characteristic factors of the user preset type service, the calculated adjusted control value is made more accurate, avoiding large deviations in the calculation results due to only considering a single element or static information.

[0130] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0131] In one embodiment, a dynamic control value control device based on data fitting is provided, and the dynamic control value control device based on data fitting corresponds to the dynamic control value control method based on data fitting in the above embodiment. Figure 3 As shown, the dynamic control value control device based on data fitting includes an acquisition module 101, a construction module 102, a calculation module 103, a matrix construction module 104, a determination module 105, a fitting module 106, a generation module 107, and an acquisition and adjustment module 108. The functional modules are described in detail as follows:

[0132] An acquisition module 101 is used to acquire characteristic factors of a user-preset service type;

[0133] A construction module 102 is configured to construct a spatiotemporal feature tensor based on the feature factors;

[0134] A calculation module 103 is used to calculate the dynamic feature range according to the spatiotemporal feature tensor;

[0135] A matrix construction module 104 is used to construct a behavior feature matrix using pre-acquired user historical behavior data;

[0136] A determination module 105 is configured to determine a feature evaluation index based on the dynamic feature range and the behavior feature matrix;

[0137] A fitting module 106 is configured to fit the feature evaluation index with an additional control value based on pre-acquired historical reference data to obtain an additional control value;

[0138] A generating module 107, configured to generate a dynamic control value based on the additional control value and a preset basic control value;

[0139] The acquisition and adjustment module 108 is configured to acquire user preset type information in real time or periodically, and adjust the dynamic control value according to the acquired user preset type information to obtain an adjusted control value.

[0140] In one embodiment, when acquiring characteristic factors of a user-preset service type, the acquisition module 101 is configured to:

[0141] Determine a user preset area according to the user preset type service, and obtain environmental feature data according to the user preset area at a set time;

[0142] Using a preset traffic system to obtain traffic condition data according to the user preset area;

[0143] Using public opinion information to obtain security dynamic data based on the user's preset area;

[0144] The environmental characteristic data, traffic condition data and safety dynamic data are aggregated into characteristic factors.

[0145] In one embodiment, when constructing the spatiotemporal feature tensor according to the feature factors, the construction module 102 is configured to:

[0146] Quantitatively processing the environmental characteristic data, traffic condition data, and safety dynamic data to obtain quantitative data;

[0147] Extracting the time dimension and space dimension of the characteristic factor, and determining the spatiotemporal dimension according to the time dimension and space dimension;

[0148] Constructing an initial spatiotemporal feature tensor according to the spatiotemporal dimension and the number of the feature factors;

[0149] The quantized data is filled into the initial spatiotemporal feature tensor to obtain the spatiotemporal feature tensor.

[0150] In one embodiment, when calculating the dynamic feature range based on the spatiotemporal feature tensor, the calculation module 103 is configured to:

[0151] Performing aggregation analysis on the spatiotemporal feature tensor to obtain aggregated spatiotemporal feature data;

[0152] Calculating the standard deviation of characteristic factors one by one according to the aggregated spatiotemporal characteristic data, and measuring the degree of dispersion of the characteristic factors by the standard deviation;

[0153] A dynamic feature range coefficient of the feature factor is set, and a dynamic feature range is determined according to the discrete degree and the dynamic feature range coefficient.

[0154] In one embodiment, when determining the feature evaluation index according to the dynamic feature range and the behavior feature matrix, the determination module 105 is configured to:

[0155] According to the behavior feature matrix, weights are assigned to the historical feature factors in the user's historical behavior data one by one to obtain historical weights corresponding to the historical feature factors;

[0156] According to the dynamic feature range, the feature factors are weighted one by one to obtain feature weights corresponding to the feature factors;

[0157] A comprehensive feature value is calculated based on the historical weight and the feature weight, and a feature evaluation index is determined based on the comprehensive feature value.

[0158] In one embodiment, when fitting the additional control value to the feature evaluation indicator based on the pre-acquired historical reference data to obtain the additional control value, the fitting module 106 is configured to:

[0159] Perform linear regression analysis on the historical reference data obtained in advance to obtain the regression coefficient;

[0160] Combining the regression coefficient with the characteristic evaluation index to obtain a combined formula;

[0161] Assigning a value to the feature evaluation index to obtain a feature value;

[0162] The combined formula is solved for an additional control value according to the regression coefficient and the eigenvalue to obtain an additional control value.

[0163] In one embodiment, when the acquisition and adjustment module 108 adjusts the dynamic control value according to the acquired user preset type information to obtain the adjusted control value, it is configured to:

[0164] Re-evaluate the feature evaluation index according to the obtained user preset type information to obtain an evaluation result;

[0165] The dynamic control value is numerically adjusted according to the evaluation result to obtain an adjusted control value.

[0166] The present invention provides a dynamic control value control device based on data fitting, which obtains characteristic factors of user preset type services for adjusting control value services, and constructs a spatiotemporal characteristic tensor according to the characteristic factors; calculates the dynamic characteristic range according to the spatiotemporal characteristic tensor; constructs a behavior characteristic matrix through pre-acquired user historical behavior data, and determines a characteristic evaluation index according to the dynamic characteristic range and the behavior characteristic matrix; performs additional control value fitting on the characteristic evaluation index according to pre-acquired historical reference data to obtain an additional control value; generates a dynamic control value according to the combination of the additional control value and a preset basic control value; obtains user preset type messages in real time or at a fixed time, and numerically adjusts the dynamic control value according to the obtained user preset type information to obtain an adjusted control value, and analyzes the characteristic factors of the user preset type services to make the calculated adjusted control value more accurate, avoiding large deviations in the calculation results caused by considering only a single element or static information.

[0167] For the specific definition of a dynamic control value control device based on data fitting, please refer to the definition of a dynamic control value control method based on data fitting above, and will not be repeated here. Each module in the above-mentioned dynamic control value control device based on data fitting can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each of the above modules.

[0168] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the server side of a dynamic control value control method based on data fitting.

[0169] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5As shown. The computer device includes a processor, memory, network interface, display screen, and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements the functions or steps on the client side of a dynamic control value control method based on data fitting.

[0170] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed:

[0171] Obtaining characteristic factors of a user-preset type of service, and constructing a spatiotemporal feature tensor based on the characteristic factors;

[0172] Calculating a dynamic feature range based on the spatiotemporal feature tensor;

[0173] Constructing a behavior feature matrix by using pre-acquired user historical behavior data, and determining a feature evaluation index according to the dynamic feature range and the behavior feature matrix;

[0174] Performing additional control value fitting on the characteristic evaluation index according to pre-acquired historical reference data to obtain an additional control value;

[0175] Generate a dynamic control value based on the combination of the additional control value and the preset basic control value;

[0176] A user preset type message is obtained in real time or periodically, and the dynamic control value is numerically adjusted according to the obtained user preset type information to obtain an adjusted control value.

[0177] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0178] Obtaining characteristic factors of a user-preset type of service, and constructing a spatiotemporal feature tensor based on the characteristic factors;

[0179] Calculating a dynamic feature range based on the spatiotemporal feature tensor;

[0180] Constructing a behavior feature matrix by using pre-acquired user historical behavior data, and determining a feature evaluation index according to the dynamic feature range and the behavior feature matrix;

[0181] Performing additional control value fitting on the characteristic evaluation index according to pre-acquired historical reference data to obtain an additional control value;

[0182] Generate a dynamic control value based on the combination of the additional control value and the preset basic control value;

[0183] A user preset type message is obtained in real time or periodically, and the dynamic control value is numerically adjusted according to the obtained user preset type information to obtain an adjusted control value.

[0184] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can be found in the relevant descriptions of the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0185] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0186] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0187] The above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. If software tools or components other than those of the company appear in the application embodiments, they are merely used for illustration and do not represent actual use. Although the present invention has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above-mentioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A dynamic control value control method based on data fitting, characterized in that: include: Obtaining characteristic factors of a user-preset type of service, and constructing a spatiotemporal feature tensor based on the characteristic factors; Calculating a dynamic feature range based on the spatiotemporal feature tensor; Constructing a behavior feature matrix by using pre-acquired user historical behavior data, and determining a feature evaluation index according to the dynamic feature range and the behavior feature matrix; Performing additional control value fitting on the characteristic evaluation index according to pre-acquired historical reference data to obtain an additional control value; Generate a dynamic control value based on the combination of the additional control value and the preset basic control value; The user preset type information is obtained in real time or periodically, and the dynamic control value is numerically adjusted according to the obtained user preset type information to obtain an adjusted control value.

2. The dynamic control value control method based on data fitting according to claim 1, characterized in that: The step of obtaining characteristic factors of a user-preset service type includes: Determine a user preset area according to the user preset type service, and obtain environmental feature data according to the user preset area at a set time; Using a preset traffic system to obtain traffic condition data according to the user preset area; Using public opinion information to obtain security dynamic data based on the user's preset area; The environmental characteristic data, traffic condition data and safety dynamic data are aggregated into characteristic factors.

3. The dynamic control value control method based on data fitting according to claim 2, characterized in that: The step of constructing a spatiotemporal feature tensor according to the feature factors includes: Quantitatively processing the environmental characteristic data, traffic condition data, and safety dynamic data to obtain quantitative data; Extracting the time dimension and space dimension of the characteristic factor, and determining the spatiotemporal dimension according to the time dimension and space dimension; Constructing an initial spatiotemporal feature tensor according to the spatiotemporal dimension and the number of the feature factors; The quantized data is filled into the initial spatiotemporal feature tensor to obtain the spatiotemporal feature tensor.

4. The dynamic control value control method based on data fitting according to claim 1, characterized in that: The calculating the dynamic feature range according to the spatiotemporal feature tensor includes: Performing aggregation analysis on the spatiotemporal feature tensor to obtain aggregated spatiotemporal feature data; Calculating the standard deviation of characteristic factors one by one according to the aggregated spatiotemporal characteristic data, and measuring the degree of dispersion of the characteristic factors by the standard deviation; A dynamic feature range coefficient of the feature factor is set, and a dynamic feature range is determined according to the discrete degree and the dynamic feature range coefficient.

5. The dynamic control value control method based on data fitting according to claim 1, characterized in that: The determining of a feature evaluation index according to the dynamic feature range and the behavior feature matrix includes: According to the behavior feature matrix, weights are assigned to the historical feature factors in the user's historical behavior data one by one to obtain historical weights corresponding to the historical feature factors; According to the dynamic feature range, the feature factors are weighted one by one to obtain feature weights corresponding to the feature factors; A comprehensive feature value is calculated based on the historical weight and the feature weight, and a feature evaluation index is determined based on the comprehensive feature value.

6. The dynamic control value control method based on data fitting according to claim 1, characterized in that: The additional control value fitting is performed on the characteristic evaluation index according to the pre-acquired historical reference data to obtain the additional control value, including: Perform linear regression analysis on the historical reference data obtained in advance to obtain the regression coefficient; Combining the regression coefficient with the characteristic evaluation index to obtain a combined formula; Assigning a value to the feature evaluation index to obtain a feature value; The combined formula is solved for an additional control value according to the regression coefficient and the eigenvalue to obtain an additional control value.

7. The dynamic control value control method based on data fitting according to claim 1, characterized in that: The step of numerically adjusting the dynamic control value according to the acquired user preset type information to obtain the adjusted control value includes: Re-evaluate the feature evaluation index according to the obtained user preset type information to obtain an evaluation result; The dynamic control value is adjusted according to the evaluation result to obtain an adjusted control value.

8. A dynamic control value control device based on data fitting, characterized in that: include: An acquisition module is used to obtain characteristic factors of a user-preset type of service; A construction module, configured to construct a spatiotemporal feature tensor according to the feature factors; A calculation module, configured to calculate a dynamic feature range based on the spatiotemporal feature tensor; A matrix construction module is used to construct a behavior feature matrix based on pre-acquired user historical behavior data; A determination module, configured to determine a feature evaluation index based on the dynamic feature range and the behavior feature matrix; A fitting module, configured to perform additional control value fitting on the characteristic evaluation index according to pre-acquired historical reference data to obtain an additional control value; A generating module, configured to generate a dynamic control value based on the additional control value and a preset basic control value; The adjustment module is used to obtain user preset type information in real time or at a fixed time, and to adjust the dynamic control value according to the obtained user preset type information to obtain an adjusted control value.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the dynamic control value control method based on data fitting are implemented as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the dynamic control value control method based on data fitting according to any one of claims 1 to 7 is implemented.

Images