Object processing method and device, equipment, storage medium and program product

Abstract

The invention provides an object processing method and device, equipment, a storage medium and a program product, and relates to the technical field of data processing.The method comprises the steps that a demand feature vector and a plurality of preset processing measures of a first object are obtained, the demand feature vector is used for representing the demand condition of the first object for resources, and the plurality of preset processing measures are used for processing the first object; the plurality of preset processing measures are processing measures for a first object; constructing a constraint condition based on the demand feature vector; generating a plurality of combinations based on the constraint condition and the plurality of preset processing measures, wherein each combination comprises at least one preset processing measure; calculating a first consumption value corresponding to each combination in the plurality of combinations; and processing the first object based on a preset processing measure included in a first combination, wherein the first combination is a combination with the lowest corresponding first consumption value in the plurality of combinations. The treatment effect of the treatment measures can be improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and specifically to an object processing method, apparatus, device, storage medium, and program product. Background Technology

[0002] Providing assistance to different groups is an important way to improve social stability. Related technologies typically require identifying specific individuals and then providing them with financial assistance. However, different groups face different problems, and using a single approach (such as financial assistance) cannot effectively address their specific issues, resulting in poor outcomes.

[0003] It is evident that the processing measures in the relevant technologies have poor processing effects. Summary of the Invention

[0004] This invention provides an object processing method, apparatus, device, storage medium, and program product to solve the problem of poor processing effect in related technologies.

[0005] To solve the above problems, the present invention is implemented as follows:

[0006] In a first aspect, embodiments of the present invention provide an object processing method, including:

[0007] Obtain the demand feature vector of the first object and multiple preset processing measures. The demand feature vector is used to characterize the resource demand of the first object, and the multiple preset processing measures are processing measures for the first object.

[0008] Construct constraints based on the aforementioned demand feature vector;

[0009] Multiple combinations are generated based on the constraints and the multiple preset processing measures, and each combination includes at least one preset processing measure.

[0010] Calculate the first consumption value corresponding to each of the plurality of combinations;

[0011] The first object is processed based on the preset processing measures included in the first combination, wherein the first combination is the combination with the lowest first consumption value among the plurality of combinations.

[0012] Secondly, embodiments of the present invention also provide an object processing apparatus, comprising:

[0013] The first acquisition module is used to acquire the demand feature vector of the first object and multiple preset processing measures. The demand feature vector is used to characterize the resource demand of the first object, and the multiple preset processing measures are processing measures for the first object.

[0014] The construction module is used to construct constraints based on the requirement feature vector;

[0015] The generation module is used to generate multiple combinations based on the constraints and the multiple preset processing measures, each combination including at least one preset processing measure;

[0016] The first calculation module is used to calculate the first consumption value corresponding to each of the plurality of combinations;

[0017] The processing module is used to process the first object based on preset processing measures included in the first combination, wherein the first combination is the combination with the lowest first consumption value among the plurality of combinations.

[0018] Thirdly, embodiments of the present invention also provide an electronic device, including a transceiver and a processor.

[0019] The transceiver is used to acquire the demand feature vector of the first object and multiple preset processing measures. The demand feature vector is used to characterize the resource demand of the first object, and the multiple preset processing measures are processing measures for the first object.

[0020] The processor is used to construct constraints based on the demand feature vector;

[0021] The processor is further configured to generate multiple combinations based on the constraints and the multiple preset processing measures, each combination including at least one preset processing measure;

[0022] The processor is further configured to calculate a first consumption value corresponding to each of the plurality of combinations;

[0023] The processor is further configured to process the first object based on preset processing measures included in the first combination, wherein the first combination is the combination with the lowest first consumption value among the plurality of combinations.

[0024] Fourthly, embodiments of the present invention provide an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the object processing method described in the first aspect.

[0025] Fifthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the object processing method described in the first aspect.

[0026] In a sixth aspect, the present invention also provides a computer program product, including computer instructions, which, when executed by a processor, implement the steps of the object processing method described in the first aspect.

[0027] In this embodiment of the invention, a demand feature vector of a first object and multiple preset processing measures are obtained. The demand feature vector characterizes the resource requirements of the first object, and the multiple preset processing measures are processing measures specifically for the first object. Constraints are constructed based on the demand feature vector. Multiple combinations are generated based on the constraints and the multiple preset processing measures, each combination including at least one preset processing measure. A first consumption value is calculated for each of the multiple combinations. The first object is processed based on the preset processing measures included in the first combination, where the first combination is the combination with the lowest corresponding first consumption value among the multiple combinations. Thus, by generating multiple combinations and calculating the first consumption value for each combination, a first combination can be determined. The first combination has the lowest first consumption value and meets the constraints, enabling the processing of the first object using the preset processing measures included in the first combination to reduce consumption while satisfying the resource requirements of the first object, thereby effectively improving the processing effect. Attached Figure Description

[0028] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a flowchart of an object processing method provided in an embodiment of the present invention;

[0030] Figure 2 This is a flowchart of the adjustment preset processing measures provided in the embodiments of the present invention;

[0031] Figure 3 This is a schematic diagram of multi-source data provided in an embodiment of the present invention;

[0032] Figure 4 This is a system structure diagram of monitoring relevant data of a first object provided in an embodiment of the present invention;

[0033] Figure 5 This is a structural diagram of an object processing device provided in an embodiment of the present invention;

[0034] Figure 6 This is a structural diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] Please see Figure 1 , Figure 1 This is a flowchart of an object processing method provided in an embodiment of the present invention, such as... Figure 1 As shown, it includes the following steps:

[0037] Step 101: Obtain the demand feature vector of the first object and multiple preset processing measures. The demand feature vector is used to characterize the resource demand of the first object, and the multiple preset processing measures are processing measures for the first object.

[0038] The aforementioned first object is the object that needs to be processed through processing measures, and the multiple preset processing measures are pre-configured measures for processing the first object. It should be noted that different objects face different problems, and their required processing measures also differ. Therefore, in this invention, the requirement feature vector of the first object is obtained, and the preset processing measures adapted to the first object are matched through the requirement feature vector.

[0039] The first target group refers to those identified as needing assistance or intervention, such as low-income individuals, which can be represented by the following formula:

[0040] ;

[0041] in the formula This represents the i-th object, in The term "i" indicates that the i-th object is the object that needs assistance or intervention, i.e., the first object.

[0042] Furthermore, the set A of multiple preset processing measures can be represented as:

[0043] ;

[0044] In the formula, a K This represents the Kth pre-set treatment measure. Different pre-set treatment measures have different contents, such as employment training measures, temporary relief measures, or housing subsidies.

[0045] The aforementioned demand feature vector is used to characterize the resource requirements of the first object. Here, "resources" refers to those that the first object needs to provide, i.e., resources that are processed through pre-defined measures to provide the resources required by the first object.

[0046] The resource requirements of the first object can be determined through the demand feature vector, which can be expressed by the following formula:

[0047] ;

[0048] q in the formula i This represents the resources required by the first object. f represents the demand feature vector. q ( ) represents a computation function.

[0049] Step 102: Construct constraints based on the aforementioned demand feature vector.

[0050] The above constraints are constraints on the preset processing measures for processing the first object. In this invention, constraints are constructed based on the demand feature vector, so that the preset processing measures for processing the first object can meet the demand corresponding to the demand feature vector of the first object, thereby achieving a better processing effect.

[0051] It should be noted that different preset processing measures may be selected simultaneously when processing the first object. However, in reality, there may be situations where several preset processing measures cannot be selected at the same time. Therefore, in this invention, while constructing constraints, different preset processing measures that can be selected simultaneously when processing the first object can also be determined according to the actual situation.

[0052] Step 103: Generate multiple combinations based on the constraints and the multiple preset processing measures, each combination including at least one preset processing measure.

[0053] Each of the aforementioned combinations includes at least one preset processing measure, and each combination meets the constraints, meaning it can satisfy the resource requirements of the first object in actual practice. The preset processing measures included in different combinations may differ; they may contain the same preset processing measure components or different preset processing measure components. For example, the combinations may include a first combination, a second combination, and a third combination. The first combination includes a first preset measure, the second combination includes a first preset measure and a second preset measure, and the third combination includes a third preset measure.

[0054] Step 104: Calculate the first consumption value corresponding to each of the multiple combinations.

[0055] It should be noted that in addition to providing resources to the first object, executing the preset processing measures also incurs certain costs, such as expenses or processing costs. Since different combinations include different preset processing measures, the cost values ​​will also differ when processing the first object using different combinations. Therefore, in this invention, after generating multiple combinations, the first cost value corresponding to each combination is calculated based on at least one preset processing measure included in each combination, thereby determining the combination with the lowest cost among the multiple combinations, thus greatly reducing the cost of processing the first object.

[0056] Step 105: Process the first object based on the preset processing measures included in the first combination, wherein the first combination is the combination with the lowest first consumption value among the plurality of combinations.

[0057] In this embodiment of the invention, a demand feature vector of a first object and multiple preset processing measures are obtained. The demand feature vector characterizes the resource requirements of the first object, and the multiple preset processing measures are processing measures specifically for the first object. Constraints are constructed based on the demand feature vector. Multiple combinations are generated based on the constraints and the multiple preset processing measures, each combination including at least one preset processing measure. A first consumption value is calculated for each of the multiple combinations. The first object is processed based on the preset processing measures included in the first combination, where the first combination is the combination with the lowest corresponding first consumption value among the multiple combinations. Thus, by generating multiple combinations and calculating the first consumption value for each combination, a first combination can be determined. The first combination has the lowest first consumption value and meets the constraints, enabling the processing of the first object using the preset processing measures included in the first combination to reduce consumption while satisfying the resource requirements of the first object, thereby effectively improving the processing effect.

[0058] In one embodiment, constructing constraints based on the demand feature vector includes:

[0059] Obtain the resource consumption and / or resource contribution vector corresponding to each of the plurality of preset processing measures;

[0060] The constraints are constructed based on the resource consumption and / or resource contribution vector.

[0061] The constraints include at least one of the following:

[0062] The sum of the resource consumption corresponding to the preset processing measures included in each combination is less than or equal to the preset total resource amount;

[0063] The sum of the resource contribution vectors corresponding to the preset processing measures included in each combination is greater than or equal to the resource demand value corresponding to the demand feature vector.

[0064] In this embodiment of the invention, by obtaining the resource consumption and / or resource contribution vector corresponding to each of the multiple preset processing measures, constraints can be constructed based on the resource consumption and / or resource contribution vector.

[0065] The resource consumption corresponding to each of the above preset processing measures can be expressed as:

[0066] ;

[0067] in the formula Indicates the preset processing measure a K The resource consumption of the Dth processing department, where the department provides the preset processing measures.

[0068] The resource contribution vector corresponding to each of the above preset processing measures can be represented as:

[0069] ;

[0070] u k Pre-set processing measure a K The corresponding resource contribution vector includes resource contributions in different dimensions, such as the employment contribution dimension. Housing contribution dimension .

[0071] The constraints can be expressed by the following formula:

[0072] ;

[0073] ;

[0074] ;

[0075] In the formula, R avail To preset the total amount of resources, u k Pre-set processing measure a K The corresponding resource contribution vector, where k is the number of preset processing measures. In a... K If the value is 1, it means that the combination includes the preset processing measure; otherwise, the combination does not include the preset processing measure.

[0076] In one embodiment, the plurality of combinations includes a second combination, and calculating the first consumption value corresponding to each of the plurality of combinations includes:

[0077] Obtain the consumption parameters corresponding to each of the plurality of preset processing measures, wherein the consumption parameters include cost consumption and / or resource consumption;

[0078] Based on the consumption parameters corresponding to the multiple preset processing measures, calculate the total cost and / or total resource scheduling value corresponding to the second combination, where the total cost is the sum of the cost consumption of the preset processing measures included in the second combination, and calculate the total resource scheduling value corresponding to the second combination, where the total resource scheduling value is the sum of the resource consumption of the preset processing measures included in the second combination.

[0079] Calculate the penalty consumption value corresponding to the second combination, where the penalty consumption value is the difference between the resource demand value corresponding to the demand feature vector and the preset demand value;

[0080] The first consumption value corresponding to the second combination is calculated based on the total cost and / or the total resource scheduling value, as well as the penalty consumption value.

[0081] In this embodiment of the invention, consumption parameters corresponding to each of the plurality of preset processing measures are obtained, the consumption parameters including cost consumption and / or resource consumption; based on the consumption parameters corresponding to the plurality of preset processing measures, the total cost value and / or total resource scheduling value corresponding to the second combination are calculated, the total cost value being the sum of the cost consumption values ​​corresponding to the preset processing measures included in the second combination; the total resource scheduling value corresponding to the second combination is calculated, the total resource scheduling value being the sum of the resource consumption values ​​corresponding to the preset processing measures included in the second combination; a penalty consumption value corresponding to the second combination is calculated, the penalty consumption value being the difference between the resource demand value corresponding to the demand feature vector and the preset demand value; based on the total cost value and / or the total resource scheduling value, and the penalty consumption value, a first consumption value corresponding to the second combination is calculated. Thus, by calculating the total cost value and / or the total resource scheduling value, and the penalty consumption value, and then calculating the first consumption value using the total cost value and / or the total resource scheduling value, and the penalty consumption value, a multi-dimensional calculation of the first consumption value is achieved, effectively improving the accuracy of the first consumption value.

[0082] In some implementations, the first cost value can be calculated simultaneously using the total cost value, the total resource scheduling value, and the penalty cost value, specifically expressed by the following formula:

[0083] ;

[0084] In the formula, L total (t) represents the first consumption value at time t, c k β and λ represent cost coefficients, respectively, and s represents weights. i(t+1) represents the resource requirement of the first object at time t+1, s target This indicates the preset requirement value.

[0085] The resource demand value is expressed by the following formula:

[0086] ;

[0087] In the formula, State() represents the calculation of the state factor.

[0088] Furthermore, the resource demand value can be updated at different times using the following formula:

[0089] ;

[0090] In the formula s i (t) represents the resource demand value at time t.

[0091] It should be noted that the above penalty cost value is defined by the following formula:

[0092] ;

[0093] In the formula, L path To penalize the consumption value, This is the time discount factor.

[0094] In some implementations, the weights of the total cost, total resource scheduling cost, and penalty consumption value can also be adjusted. Specifically, the adjustment process can be represented by the following formula:

[0095] ;

[0096] By adjusting the above formula, in L aug The weights are determined to minimize the minimum value. In the formula... These are Lagrange multipliers used to reflect resource constraint pressures. The preset total amount of resources at time t.

[0097] In the above formula Specifically, it is calculated using the following formula:

[0098] ;

[0099] In the formula, x is a coefficient.

[0100] In this way, the first consumption value for different combinations can be calculated using the above formula, thereby achieving the optimal preset processing measures and realizing a resource allocation strategy that prioritizes proximity and low time consumption. Simultaneously, scheduling constraints, such as budget limitations, response capabilities, and geographical accessibility, are considered to ensure that the resource allocation process is feasible, reasonable, and efficient.

[0101] In one embodiment, the method further includes:

[0102] Obtain the behavior trend vector of the first object, which is used to characterize the changes in the behavior data of the first object after processing;

[0103] Calculate the change value of the first object based on the behavioral trend vector;

[0104] If the change value is greater than a set change threshold, the preset processing measures included in the first combination are adjusted.

[0105] In this embodiment of the invention, a behavior trend vector of the first object is obtained, which characterizes the changes in the behavior data of the first object after processing; a change value of the first object is calculated based on the behavior trend vector; and if the change value is greater than a set change threshold, the preset processing measures included in the first combination are adjusted. Thus, by adjusting the preset processing measures for the first object according to the changes in its behavior data, the first object can be continuously and effectively processed, further improving the processing effect of the preset processing measures.

[0106] Specifically, such as Figure 2 As shown, for the first object, in the process of handling it using the preset handling measures, it is also necessary to analyze its abnormal behavior and generate an early warning level so as to facilitate statistical analysis and adjust the preset handling measures.

[0107] In some implementations, the change value of the first object is calculated based on the behavioral trend vector, which can be specifically expressed by the following formula:

[0108] ;

[0109] ;

[0110] In the formula, D M (E i (t) represents the change value. Let W represent the behavior trend vector, W represent the time window length, and vec() represents flattening the multidimensional behavior trend vector. and These represent the mean and covariance of the global behavioral trend, respectively.

[0111] Furthermore, the aforementioned threshold for change is set. It can be calculated using the following formula:

[0112] ;

[0113] in the formula The mean value representing the overall behavioral trend. Variance representing the overall behavioral trend. This represents the weighting coefficient.

[0114] In one embodiment, the first combination includes at least one first measure, and the preset processing measure for adjusting the first combination includes:

[0115] Calculate the warning level based on the changes;

[0116] Obtain the adjustment threshold corresponding to each first measure;

[0117] The second measure is adjusted, wherein the adjustment threshold corresponding to the at least one first measure is less than or equal to the warning level.

[0118] In this embodiment of the invention, an early warning level is calculated based on the change value; an adjustment threshold corresponding to each first measure is obtained; and a second measure is adjusted, wherein the second measure is a measure in the at least one first measure whose corresponding adjustment threshold is less than or equal to the early warning level. Thus, by calculating the early warning level and obtaining the adjustment threshold corresponding to each first measure, it is determined whether the first measures need to be adjusted; and by adjusting the second measure in the at least one first measure, the adjusted measure can meet the situation after the change in the first object's behavioral data, further improving the processing effect.

[0119] In some implementations, the warning level is calculated based on the change value, which can be specifically expressed by the following formula:

[0120] ;

[0121] In the formula, α1, α2, and α3 are normalized weighting coefficients, and H(F) i (t) represents the discrete entropy of the behavior distribution, through... The calculated value, p, is used to reflect the instability of the behavior. j Let be the probability of action j occurring. This is for scheduling residuals.

[0122] Furthermore, multiple levels can be set, along with a corresponding numerical range for each level, through calculation. Determine the warning level. This can be expressed using the following formula:

[0123] ;

[0124] in the formula , and These represent extreme values ​​at different levels.

[0125] It should be noted that, because different measures are implemented by different entities and address different specific situations, adjustments to the measures need to be determined based on the warning level. For example, educational assistance measures have a higher threshold and are applicable to more scenarios; while economic assistance measures have a lower threshold and can be discontinued once the recipient has reached a certain income. Specifically, this invention sets an adjustment threshold for each first measure, and the need for adjustment to the first measure is determined by adjusting the threshold.

[0126] In some implementations, determining whether the first measure needs adjustment by adjusting the threshold can be expressed by the following formula:

[0127] ;

[0128] w in the formula d As weight, Let I be the adjustment threshold corresponding to the first measure, and let I() be the indicator function. R is obtained through calculation. i The value of (t)[d] determines whether the first measure needs to be adjusted.

[0129] In some implementations, adjusting the second measure can be achieved by pre-configuring supplementary measures (such as providing food, providing consultation, etc.) for each second measure, and adjusting the second measure using these supplementary measures when necessary. This can be specifically expressed by the following formula:

[0130] ;

[0131] in the formula This indicates the second measure after the adjustment. This indicates the second measure prior to the adjustment. This indicates supplementary measures.

[0132] In one embodiment, before obtaining the first requirement feature vector of the first object and multiple preset processing measures, the method further includes:

[0133] Retrieve multi-source data corresponding to each of the multiple initial objects;

[0134] Based on the multi-source data corresponding to each initial object, extract the corresponding data feature vector of the initial object;

[0135] Predict the probability of each initial object based on the data feature vectors corresponding to the plurality of initial objects;

[0136] The first object is determined, which is the initial object among the plurality of initial objects whose probability is greater than a first set threshold.

[0137] In this embodiment of the invention, multi-source data corresponding to each of a plurality of initial objects is acquired; data feature vectors corresponding to each initial object are extracted based on the multi-source data corresponding to each initial object; the probability of each initial object is predicted based on the data feature vectors corresponding to the plurality of initial objects; and the first object is determined, wherein the first object is the initial object among the plurality of initial objects whose probability is greater than a first preset threshold. Thus, by separately determining the probability of each initial object, it is determined whether the initial object is the first object, and then a corresponding preset processing measure is determined for the first object.

[0138] In some implementations, the multi-source data corresponding to each initial object may include data from operators and data from different implementing departments. Specifically, such as... Figure 3 As shown, multi-source data includes data from multiple departments. The initial dataset of multi-source data can be represented as: D = {D} tel D gov}, D tel For data from operators, such as location tracking, communication behavior (call frequency, base station handover), data consumption, and bill payment behavior; D gov Data for different implementing departments, such as family structure, disability level, household registration, low-income records, housing information, etc.

[0139] Furthermore, the collected multi-source data can be encrypted using a combination of a high-entropy grouping perturbation mechanism based on quantum random numbers and homomorphic encryption.

[0140] Furthermore, the data feature vector can be represented as ;x i Let be the data feature vector of the i-th initial object. A high-entropy seed-driven adaptive perturbation coding mechanism is introduced for each initial object i's multi-source data fragment. We introduce a hybrid index of Shannon entropy and structural complexity, specifically expressed by the following formula:

[0141] ;

[0142] ;

[0143] In the formula, p(x) i,j Let be the probability that the j-th data segment is represented by , and let PCA be the probability of the j-th data k ( ) is principal component analysis, Var() is used to calculate variance, and Rank() is used to rank the data segments. is a coefficient.

[0144] Let the disturbance intensity control function for:

[0145] ;

[0146] Then introduce the disturbance control matrix:

[0147] ;

[0148] in From a quantum random source, By controlling the perturbation amplitude, sensitive fields can be better protected for privacy. Features after scrambling. Homomorphic encryption .

[0149] Furthermore, to support verifiable encrypted collaborative analysis across multiple execution departments, a multi-level homomorphic nested structure is introduced, combined with a multi-key collaborative decryption protocol: Original encryption process (multi-party joint modeling scenario): [The following appears to be a separate, unrelated section:] User feature vector fragments... Multi-layered encryption is performed as follows:

[0150] ;

[0151] in the formula For the final encrypted data, the first level PK1 is hosted by the telecommunications operator; the second level PK2 is hosted by the department; and the third level PK3 is managed uniformly by the main scheduling platform. This nested structure ensures that no single party can decrypt the joint data; it supports multi-party joint homomorphic computation; and decryption requires collaborative completion through a multi-party secure computation protocol.

[0152] Furthermore, based on the encrypted data after collaborative data access, a high-dimensional tensor representation for heterogeneous data from multiple departments is designed to characterize individual states, temporal evolution, and correlation features. The data structure suffers from issues such as asynchronous acquisition, different field names, semantic inconsistencies, and inconsistent time alignment granularity, requiring unified modeling.

[0153] The characteristic data (ciphertext form) of each initial object i in different execution departments s are as follows: Define the cross-departmental feature splicing tensor as:

[0154] ;

[0155] Where S represents the number of departments accessing the system; The ciphertext feature concatenation operation remains in a homomorphically computeable state. Considering time evolution and multi-layered structure, a third-order tensor is constructed:

[0156] ;

[0157] Where F i Represents the latent factor vector (extracted from cross-source synthesis factors through secure decomposition). This represents the temporal feature index. The specific tensor expansion form is:

[0158] ;

[0159] in A cross-departmental feature-to-factor mapping matrix; ψ j,q For feature-to-behavior pattern space mapping; τ j,r It is a time-series index weight, supporting dynamic windowing processing.

[0160] In some implementations, the probability of each initial object is predicted based on the data feature vectors corresponding to the plurality of initial objects. This can be achieved by predicting the sub-probability of an initial object using multiple prediction models, and then weighting the sub-probabilities corresponding to each prediction model to obtain the probability of the initial object.

[0161] Among them, several prediction models include the LightGBM prediction model, the XGBoost prediction model, and the CatBoost prediction model.

[0162] For example, for the LightGBM prediction model, parameters such as learning rate, boosting type, and n_estimators can be set, and the tree depth can be set for maxdepth and numleaves to prevent overfitting. The GridSearchCV function from sklearn can also be used for grid search.

[0163] The parameters of the LightGBM prediction model can be set as shown in the table below:

[0164]

[0165] The parameters of the LightGBM prediction model are set using the parameters in the table above, so that the probability of the initial object can be predicted using the LightGBM prediction model.

[0166] Furthermore, for the XGBoost prediction model, similar to building the LightGBM model, some parameters need to be adjusted after the XGBoost model is built to make it more effective. Again, a higher learning rate is initially chosen, and the GridSearchCV function from sklearn is used to perform a grid search to determine the number of trees. Then, given the learning rate and the number of trees, the decision trees are fine-tuned, followed by regularization tuning, and finally, the learning rate is reduced to determine the final parameters.

[0167] The parameters of the XGBoost prediction model can be set as shown in the table below:

[0168]

[0169] The parameters of the XGBoost prediction model are set using the parameters in the table above, so that the probability of the initial object can be predicted using the XGBoost prediction model.

[0170] Furthermore, for the CatBoost prediction model, a grid search method was used to find the ideal number of decision trees. The most significant feature of the CatBoost prediction model is its ability to automatically perform one-hot encoding of specific attributes, while also achieving high model quality without parameter tuning. Therefore, after several simple model runs, the final model parameters were determined.

[0171] The parameters of the CatBoost prediction model can be set as shown in the table below:

[0172]

[0173] The parameters of the CatBoost prediction model are set using the parameters in the table above, so that the probability of the initial object can be predicted using the CatBoost prediction model.

[0174] In some implementations, the Voting model fusion method can be used to fuse the prediction results of different prediction models. Specifically, the calculation process can be represented by the following formula:

[0175] M voting =n1M1+n2M2+⋯+n k M k ;

[0176] In the formula, M voting The probabilities of the initial objects, M1 to M k The subprobabilities predicted by different models, n1 to n k These are the weighting coefficients. Where n1 + n2 + ... + n k =1.

[0177] The weighting coefficients can be determined using the following formula:

[0178] ;

[0179] Where: the true label is y, the predicted probability given by the model is p, and log represents the natural logarithm. When y=1, the first term ylog(p) takes effect, representing the logarithm of the probability that the model predicts a positive example when the actual example is positive; when y=0, the second term (1-y)log(1-p) takes effect, representing the logarithm of the probability that the model predicts a negative example when the actual example is negative.

[0180] In some implementations, after processing the first object based on the preset processing measures included in the first combination, the relevant data of the first object can also be monitored to determine the processing effect on the first object.

[0181] Furthermore, such as Figure 4 As shown, a multi-source data status scheduling and monitoring center can be constructed to achieve visualization, distributed linkage, and stable operation. Specifically, the encrypted and decoded demand feature vectors, the first object identification results, the preset processing measures for the first object, and the behavior anomaly scores can all be integrated into the scheduling control node. Preset processing measures across departments can be broken down and tasks can be assigned. Resource scheduling, unit matching, and response action distribution can be performed for different preset processing parameters in the first combination. Scheduling status feedback and system monitoring include tracking the response status (success / failure / waiting), processing latency, and whether anomalies are triggered for each preset processing measure through message queues (such as Kafka) or event streams (such as Flink) for status tracking and system situation assessment.

[0182] In addition, data can be recorded through logs. A chain-structured log recording mechanism can be used, including: a path generation log, which records different combinations generated from scheduling inference and their corresponding strategy sources and resource availability assessments; a behavior anomaly scoring log, which saves the evolution trajectory of multi-scale trend scores and can be used for historical backtracking and responsibility attribution; and a scheduling response log, used to audit whether the responses of each node in the scheduling path are compliant and timely, and whether there are problems such as "idling" or "dead nodes".

[0183] In addition, a continuous learning and self-deployment mechanism for the model can be set up. Specifically, model loading and version scheduling, new identification models (such as low-income identification), scheduling generators (path inference engines), and model performance are dynamically weighted and evaluated through micro-indicator feedback scores (such as AUC, F1, path hit rate), and the best are retained; a multi-department deployment plug-in mechanism supports plug-in deployment between different execution departments, with decentralized data control and centralized unified access permissions, and supports private deployment and hybrid cloud integration.

[0184] Please see Figure 5 , Figure 5 This is a structural diagram of an object processing device provided in an embodiment of the present invention, such as... Figure 5 As shown, the object processing apparatus 500 includes:

[0185] The first acquisition module 501 is used to acquire the demand feature vector of the first object and multiple preset processing measures. The demand feature vector is used to characterize the resource demand of the first object, and the multiple preset processing measures are processing measures for the first object.

[0186] Construction module 502 is used to construct constraints based on the requirement feature vector;

[0187] The generation module 503 is used to generate multiple combinations based on the constraints and the multiple preset processing measures, each combination including at least one preset processing measure;

[0188] The first calculation module 504 is used to calculate the first consumption value corresponding to each of the plurality of combinations;

[0189] The processing module 505 is used to process the first object based on the preset processing measures included in the first combination, wherein the first combination is the combination with the lowest first consumption value among the plurality of combinations.

[0190] In one embodiment, the building module 502 includes:

[0191] The first acquisition unit is used to acquire the resource consumption and / or resource contribution vector corresponding to each of the plurality of preset processing measures;

[0192] The construction unit is used to construct the constraints based on the resource consumption and / or resource contribution vector;

[0193] The constraints include at least one of the following:

[0194] The sum of the resource consumption corresponding to the preset processing measures included in each combination is less than or equal to the preset total resource amount;

[0195] The sum of the resource contribution vectors corresponding to the preset processing measures included in each combination is greater than or equal to the resource demand value corresponding to the demand feature vector.

[0196] In one embodiment, the plurality of combinations includes a second combination, and the first computing module 504 includes:

[0197] The second acquisition unit is used to acquire the consumption parameters corresponding to each of the plurality of preset processing measures, wherein the consumption parameters include cost consumption and / or resource consumption.

[0198] The first calculation unit is used to calculate the total cost and / or total resource scheduling value corresponding to the second combination based on the consumption parameters corresponding to the plurality of preset processing measures, wherein the total cost is the sum of the cost consumption of the preset processing measures included in the second combination, and to calculate the total resource scheduling value corresponding to the second combination, wherein the total resource scheduling value is the sum of the resource consumption of the preset processing measures included in the second combination.

[0199] The second calculation unit is used to calculate the penalty consumption value corresponding to the second combination, wherein the penalty consumption value is the difference between the resource demand value corresponding to the demand feature vector and the preset demand value;

[0200] The third calculation unit is used to calculate the first consumption value corresponding to the second combination based on the total cost and / or the total resource scheduling value, and the penalty consumption value.

[0201] In one embodiment, the object processing apparatus 500 further includes:

[0202] The second acquisition module is used to acquire the behavior trend vector of the first object, and the behavior trend vector is used to characterize the change of the behavior data of the first object after processing.

[0203] The second calculation module is used to calculate the change value of the first object based on the behavior trend vector;

[0204] An adjustment module is used to adjust the preset processing measures included in the first combination when the change value is greater than a set change threshold.

[0205] In one embodiment, the first combination includes at least one first measure, and the adjustment module includes:

[0206] The fourth calculation unit is used to calculate the warning level based on the changed value;

[0207] The third acquisition unit is used to acquire the adjustment threshold corresponding to each first measure;

[0208] An adjustment unit is used to adjust a second measure, wherein the second measure is a measure in which the adjustment threshold corresponding to the at least one first measure is less than or equal to the warning level.

[0209] In one embodiment, the object processing apparatus 500 further includes:

[0210] The third acquisition module is used to acquire multi-source data corresponding to each of the multiple initial objects;

[0211] The extraction module is used to extract the data feature vector corresponding to each initial object based on the multi-source data corresponding to each initial object;

[0212] The prediction module is used to predict the probability of each initial object based on the data feature vectors corresponding to the plurality of initial objects;

[0213] The determining module is used to determine the first object, wherein the first object is an initial object among the plurality of initial objects with a probability greater than a first set threshold.

[0214] The object processing apparatus provided in this embodiment of the invention can implement each process of each embodiment of the above-described object processing method. The technical features are one-to-one and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0215] It should be noted that the object processing device in the embodiments of the present invention can be a device, or it can be a component, integrated circuit, or chip in an electronic device.

[0216] This invention also provides an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the above-described functionality. Figure 1 The various processes of the object processing method embodiments shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0217] For details, see Figure 6 As shown, this embodiment of the invention also provides an electronic device, including a bus 601, a transceiver 602, an antenna 603, a bus interface 604, a processor 605, and a memory 606.

[0218] The transceiver 602 is used to acquire the demand feature vector of the first object and multiple preset processing measures. The demand feature vector is used to characterize the resource demand of the first object, and the multiple preset processing measures are processing measures for the first object.

[0219] The processor 605 is used to construct constraints based on the demand feature vector;

[0220] The processor 605 is further configured to generate multiple combinations based on the constraints and the multiple preset processing measures, each combination including at least one preset processing measure;

[0221] The processor 605 is further configured to calculate a first consumption value corresponding to each of the plurality of combinations;

[0222] The processor 605 is further configured to process the first object based on preset processing measures included in the first combination, wherein the first combination is the combination with the lowest first consumption value among the plurality of combinations.

[0223] In one embodiment, constructing constraints based on the demand feature vector includes:

[0224] Obtain the resource consumption and / or resource contribution vector corresponding to each of the plurality of preset processing measures;

[0225] The constraints are constructed based on the resource consumption and / or resource contribution vector.

[0226] The constraints include at least one of the following:

[0227] The sum of the resource consumption corresponding to the preset processing measures included in each combination is less than or equal to the preset total resource amount;

[0228] The sum of the resource contribution vectors corresponding to the preset processing measures included in each combination is greater than or equal to the resource demand value corresponding to the demand feature vector.

[0229] In one embodiment, the plurality of combinations includes a second combination, and calculating the first consumption value corresponding to each of the plurality of combinations includes:

[0230] Obtain the consumption parameters corresponding to each of the plurality of preset processing measures, wherein the consumption parameters include cost consumption and / or resource consumption;

[0231] Based on the consumption parameters corresponding to the multiple preset processing measures, calculate the total cost and / or total resource scheduling value corresponding to the second combination, where the total cost is the sum of the cost consumption of the preset processing measures included in the second combination, and calculate the total resource scheduling value corresponding to the second combination, where the total resource scheduling value is the sum of the resource consumption of the preset processing measures included in the second combination.

[0232] Calculate the penalty consumption value corresponding to the second combination, where the penalty consumption value is the difference between the resource demand value corresponding to the demand feature vector and the preset demand value;

[0233] The first consumption value corresponding to the second combination is calculated based on the total cost and / or the total resource scheduling value, as well as the penalty consumption value.

[0234] In one embodiment, the transceiver 602 is further configured to acquire a behavior trend vector of the first object, the behavior trend vector being used to characterize the changes in the behavior data of the first object after processing;

[0235] The processor 605 is further configured to calculate the change value of the first object based on the behavior trend vector;

[0236] The processor 605 is further configured to adjust the preset processing measures included in the first combination when the change value is greater than a set change threshold.

[0237] In one embodiment, the first combination includes at least one first measure, and the preset processing measure for adjusting the first combination includes:

[0238] Calculate the warning level based on the changes;

[0239] Obtain the adjustment threshold corresponding to each first measure;

[0240] The second measure is adjusted, wherein the adjustment threshold corresponding to the at least one first measure is less than or equal to the warning level.

[0241] In one embodiment, the transceiver 602 is further configured to acquire multi-source data corresponding to each of the multiple initial objects;

[0242] The processor 605 is further configured to extract the data feature vector corresponding to each initial object based on the multi-source data corresponding to each initial object;

[0243] The processor 605 is further configured to predict the probability of each initial object based on the data feature vectors corresponding to the plurality of initial objects;

[0244] The processor 605 is further configured to determine the first object, wherein the first object is an initial object among the plurality of initial objects with a probability greater than a first set threshold.

[0245] exist Figure 6 In this document, a bus architecture (represented by bus 601) is used. Bus 601 may include any number of interconnected buses and bridges, linking various circuits including one or more processors represented by processor 605 and memory represented by memory 606. Bus 601 may also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 604 provides an interface between bus 601 and transceiver 602. Transceiver 602 may be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 605 is transmitted over a wireless medium via antenna 603, which further receives data and transmits data to processor 605.

[0246] Processor 605 manages bus 601 and general processing, and also provides various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. Memory 606 can be used to store data used by processor 605 during operation.

[0247] Optionally, the processor 605 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a graphics processing unit (GPU).

[0248] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the above-described functions. Figure 1 The various processes of the corresponding object processing method embodiments, which can achieve the same technical effect, will not be described again here to avoid repetition. The computer-readable storage medium mentioned includes, for example, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0249] The present invention also provides a computer program product, including computer instructions that, when executed by a processor, implement the above-described... Figure 1 The various processes of the corresponding object processing method embodiments can achieve the same technical effect, and will not be described again here to avoid repetition.

[0250] In the embodiments of this invention, the terms "first," "second," etc., are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices. Additionally, the use of "and / or" in this application indicates at least one of the connected objects, such as A and / or B and / or C, representing eight possibilities: A alone, B alone, C alone, both A and B present, both B and C present, both A and C present, and A, B, and C present.

[0251] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0252] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or second terminal device, etc.) to execute the methods of the various embodiments of this application.

[0253] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. An object processing method, characterized in that, include: Obtain the demand feature vector of the first object and multiple preset processing measures. The demand feature vector is used to characterize the resource demand of the first object, and the multiple preset processing measures are processing measures for the first object. Construct constraints based on the aforementioned demand feature vector; Multiple combinations are generated based on the constraints and the multiple preset processing measures, and each combination includes at least one preset processing measure. Calculate the first consumption value corresponding to each of the plurality of combinations; The first object is processed based on the preset processing measures included in the first combination, wherein the first combination is the combination with the lowest first consumption value among the plurality of combinations.

2. The method as described in claim 1, characterized in that, The construction of constraints based on the demand feature vector includes: Obtain the resource consumption and / or resource contribution vector corresponding to each of the plurality of preset processing measures; The constraints are constructed based on the resource consumption and / or resource contribution vector. The constraints include at least one of the following: The sum of the resource consumption corresponding to the preset processing measures included in each combination is less than or equal to the preset total resource amount; The sum of the resource contribution vectors corresponding to the preset processing measures included in each combination is greater than or equal to the resource demand value corresponding to the demand feature vector.

3. The method as described in claim 1, characterized in that, The plurality of combinations includes a second combination, and the calculation of the first consumption value corresponding to each of the plurality of combinations includes: Obtain the consumption parameters corresponding to each of the plurality of preset processing measures, wherein the consumption parameters include cost consumption and / or resource consumption; Based on the consumption parameters corresponding to the multiple preset processing measures, calculate the total cost and / or total resource scheduling value corresponding to the second combination, where the total cost is the sum of the cost consumption of the preset processing measures included in the second combination, and calculate the total resource scheduling value corresponding to the second combination, where the total resource scheduling value is the sum of the resource consumption of the preset processing measures included in the second combination. Calculate the penalty consumption value corresponding to the second combination, where the penalty consumption value is the difference between the resource demand value corresponding to the demand feature vector and the preset demand value; The first consumption value corresponding to the second combination is calculated based on the total cost and / or the total resource scheduling value, as well as the penalty consumption value.

4. The method according to any one of claims 1 to 3, characterized in that, The method further includes: Obtain the behavior trend vector of the first object, which is used to characterize the changes in the behavior data of the first object after processing; Calculate the change value of the first object based on the behavioral trend vector; If the change value is greater than a set change threshold, the preset processing measures included in the first combination are adjusted.

5. The method as described in claim 4, characterized in that, The first combination includes at least one first measure, and the preset processing measure for adjusting the first combination includes: Calculate the warning level based on the changes; Obtain the adjustment threshold corresponding to each first measure; The second measure is adjusted, wherein the adjustment threshold corresponding to the at least one first measure is less than or equal to the warning level.

6. The method according to any one of claims 1 to 3, characterized in that, Before obtaining the first requirement feature vector of the first object and multiple preset processing measures, the method further includes: Retrieve multi-source data corresponding to each of the multiple initial objects; Based on the multi-source data corresponding to each initial object, extract the corresponding data feature vector of the initial object; Predict the probability of each initial object based on the data feature vectors corresponding to the plurality of initial objects; The first object is determined, which is the initial object among the plurality of initial objects whose probability is greater than a first set threshold.

7. An object processing apparatus, characterized in that, include: The first acquisition module is used to acquire the demand feature vector of the first object and multiple preset processing measures. The demand feature vector is used to characterize the resource demand of the first object, and the multiple preset processing measures are processing measures for the first object. The construction module is used to construct constraints based on the requirement feature vector; The generation module is used to generate multiple combinations based on the constraints and the multiple preset processing measures, each combination including at least one preset processing measure; The first calculation module is used to calculate the first consumption value corresponding to each of the plurality of combinations; The processing module is used to process the first object based on preset processing measures included in the first combination, wherein the first combination is the combination with the lowest first consumption value among the plurality of combinations.

8. An electronic device, characterized in that, Including transceivers and processors, The transceiver is used to acquire the demand feature vector of the first object and multiple preset processing measures. The demand feature vector is used to characterize the resource demand of the first object, and the multiple preset processing measures are processing measures for the first object. The processor is used to construct constraints based on the demand feature vector; The processor is further configured to generate multiple combinations based on the constraints and the multiple preset processing measures, each combination including at least one preset processing measure; The processor is further configured to calculate a first consumption value corresponding to each of the plurality of combinations; The processor is further configured to process the first object based on preset processing measures included in the first combination, wherein the first combination is the combination with the lowest first consumption value among the plurality of combinations.

9. An electronic device, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the object processing method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the object processing method as described in any one of claims 1 to 7.

11. A computer program product, characterized in that, It includes computer instructions that, when executed by a processor, implement the steps of the object processing method as described in any one of claims 1 to 7.

Images