Weighted random expert extraction method and system based on A-Res algorithm

Abstract

The invention discloses a weighted random expert extraction method and system based on an A-Res algorithm, and relates to the technical field of data processing and random sampling, and the system comprises a database which is used for collecting expert information, storing the expert information, providing expert information retrieval service, and maintaining the stored expert information; the extraction unit is used for acquiring the sample set from the database, setting extraction conditions, and extracting from the sample set based on the extraction conditions to obtain an expert extraction result; the statistical unit is used for counting expert extraction results obtained by the extraction unit; according to the system, through a dual mechanism of A-Res preprocessing probability weight and sample set random sorting, on the premise of not significantly increasing algorithm complexity, code implementation complexity is reduced, interference of pseudo-random numbers is reduced, and independence and fairness of sampling results are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing and random sampling, and specifically, to a weighted random expert extraction method and system based on the A-Res algorithm. Background Art

[0002] When sampling a weighted sample set with reward and punishment interest relationships for targets such as review experts, agency organizations, framework agreement manufacturers, etc., there are currently the following two sampling methods:

[0003] I. Sampling method based on the reservoir algorithm:

[0004] The reservoir algorithm is a classic algorithm for randomly selecting a fixed-size sample in a data stream. It maintains a reservoir of a fixed size and dynamically updates the sample through a probability mechanism to ensure that the probability of each new sample being selected is equal. It can be proved by mathematical induction that when randomly extracting k elements from a sample set of size n, using the reservoir algorithm can ensure that the probability of each element being drawn is k / n, and since this algorithm only traverses the sample set once, the algorithm complexity can be guaranteed to be only O(n).

[0005] II. Sampling method based on the A-Res algorithm:

[0006] The A-Res algorithm is a weighted random sampling algorithm proposed by Vitter in the paper "Weighted Random Sampling with a Reservoir" in 2006. Based on the reservoir algorithm, it assigns a weight to each element to ensure that the probability of each element being drawn is proportional to its weight. Its core idea is to dynamically maintain the elements in the reservoir through the mechanism of retention probability and virtual "water level line". Since each time an element is drawn, it competes with k elements in the reservoir for the retention probability, the algorithm complexity is slightly increased to O(n )

[0007] The following problems exist in the code implementation of traditional sample set random extraction methods:

[0008] First, if only the sampling method based on the random() function is simply used, since the random numbers generated by the computer are pseudo-random numbers, it is difficult to ensure the fairness and independence of each extraction result, and for the need of weighted random extraction, especially when the weights vary greatly, its implementation method is relatively complex;

[0009] Second, if the traditional sampling method based on the A-Res algorithm is adopted, after reaching the virtual water level line, the retention probability of each subsequent sample needs to be calculated separately, which includes the sorting of the probability weights of all samples in the reservoir and the comparison with the probability weights of subsequent samples. If the extraction quantities of different sample sets are different, it increases the complexity of code implementation. Moreover, the generation of sample probability weights is also based on the pseudo-random numbers generated by the computer. If the sample sequence remains unchanged each time, it is easily interfered by the reproducibility and correlation of the pseudo-random numbers, thus affecting the fairness and independence of the extraction results. Summary of the Invention

[0010] The object of the present invention is to simply and efficiently implement expert sampling and reduce the interference of the reproducibility and correlation of pseudo-random numbers.

[0011] To achieve the above object of the invention, the present invention provides a weighted random expert extraction method based on the A-Res algorithm, and the method includes:

[0012] Step 1: Collect and store expert information;

[0013] Step 2: Obtain a sample set based on the stored expert information, set extraction conditions, and obtain an expert extraction result by extracting from the sample set based on the extraction conditions;

[0014] Step 3: Statistically analyze the expert extraction results obtained by the extraction unit;

[0015] Among them, obtaining the expert extraction result by extracting from the sample set based on the extraction conditions specifically includes:

[0016] Obtain a first sample set;

[0017] Randomly arrange the first sample set to obtain a second sample set;

[0018] Calculate the total weight of all samples in the second sample set;

[0019] Each sample in the second sample set generates a random base number respectively;

[0020] Based on the total weight of all samples, the random base number of each sample, and the sample weight of each sample, calculate the first probability weight corresponding to each sample respectively;

[0021] Update the first probability weight of each sample, and each sample obtains the corresponding second probability weight respectively;

[0022] Based on the second probability weight of each sample, sort them in descending order of the second probability weight to generate an extraction result sequence;

[0023] Select the first several samples from the extraction result sequence to obtain the expert extraction result.

[0024] Among them, in this method, the sample weights and random numbers are combined through the A-Res preprocessing algorithm to generate the second probability weights, and the interference of pseudo-random numbers is reduced through random sorting before sampling, so as to achieve efficient and fair weighted random extraction.

[0025] Preferably, step 1 further includes: maintaining the stored expert information, specifically including:

[0026] Adding, deleting, and changing the stored expert information, and maintaining the extraction weights of the expert information.

[0027] Preferably, the first sample set is S, and the expression of S is:

[0028] ;

[0029] Among them, is the ID value of the m-th sample in the first sample set, 1 ≤ m ≤ n, where n is the total number of samples in the first sample set, is the extraction weight of the m-th sample in the first sample set.

[0030] Preferably, the calculation method of the sum of the weights of all samples is:

[0031] ;

[0032] Among them, W is the sum of the weights of all samples, is the extraction weight of the i-th sample in the second sample set, 1 ≤ i ≤ n, where n is the total number of samples in the first sample set.

[0033] Preferably, the random base number of each sample in the second sample set is generated in the following manner:

[0034] ;

[0035] Among them, is the random base number of each sample in the second sample set, 1 ≤ i ≤ n, where n is the total number of samples in the first sample set, The function represents taking a random value between 0 and 1.

[0036] Preferably, the calculation method of the first probability weight is:

[0037] ;

[0038] Among them, is the first probability weight of the i-th sample in the second sample set, W is the sum of the weights of all samples, is the extraction weight of the i-th sample in the second sample set, is the random base number for each sample in the second sample set, where 1 ≤ i ≤ n and n is the total number of samples in the first sample set.

[0039] Preferably, the calculation method of the second probability weight is:

[0040] ;

[0041] Among them, is the second probability weight of the i-th sample in the second sample set, W is the total weight of all samples, is the extraction weight of the i-th sample in the second sample set, is the random base number for each sample in the second sample set, where 1 ≤ i ≤ n and n is the total number of samples in the first sample set.

[0042] Preferably, the extraction result sequence is R, and the expression of R is:

[0043] ;

[0044] Among them, is the ID value of the i-th sample in the second sample set, is the second probability weight of the i-th sample in the second sample set, is the second probability weight of the (i - 1)-th sample in the second sample set, is the second probability weight of the (i + 1)-th sample in the second sample set, W is the total weight of all samples, where 1 ≤ i ≤ n and n is the total number of samples in the first sample set.

[0045] The present invention also provides a weighted random expert extraction system based on the A-Res algorithm, and the system includes:

[0046] A database for collecting expert information, storing expert information, providing expert information retrieval services, and maintaining the stored expert information;

[0047] An extraction unit for obtaining a sample set from the database, setting extraction conditions, and extracting an expert extraction result from the sample set based on the extraction conditions;

[0048] A statistical unit for statistically analyzing the expert extraction results obtained by the extraction unit;

[0049] Among them, the method for obtaining the expert extraction result is:

[0050] Obtain the first sample set;

[0051] Randomly arrange the first sample to obtain the second sample set;

[0052] Calculate the total weight of all samples in the second sample set;

[0053] Each sample in the second sample set generates a random base number respectively;

[0054] Based on the total weight of all samples, the random base number of each sample, and the sample weight of each sample, calculate the first probability weight value corresponding to each sample respectively;

[0055] Update the first probability weight value of each sample, and each sample obtains the corresponding second probability weight value;

[0056] Based on the second probability weight value of each sample, sort them in descending order of the second probability weight value to generate an extraction result sequence;

[0057] Select the first several samples from the extraction result sequence to obtain the expert extraction result.

[0058] Preferably, the system further includes a basic configuration unit for performing basic configuration and settings on the system.

[0059] One or more technical solutions provided by the present invention have at least the following technical effects or advantages:

[0060] 1. Simple code implementation: During the extraction process, for the processing of each sample, only its probability weight value needs to be generated and saved, without comparing with the reservoir, so there is no need to maintain the virtual water level line of the traditional A-Res algorithm. After traversing the sample set, perform a quick sort on the sample set again (the algorithm complexity is O(n )), and delimit the virtual water level line according to the number of samples to be extracted, so the code implementation is more simple, and the algorithm complexity corresponding to the system only increases at a constant level.

[0061] 2. Fairness guarantee: The A-Res algorithm idea adopted in this system can strictly ensure that the probability of each sample being drawn is proportional to its weight, and the logarithm is used to revise the first probability weight value, which does not affect the extraction result while ensuring the calculation accuracy. Moreover, the sample set is randomly sorted in the preprocessing stage of each extraction to ensure the independence of the initial order of each extraction, and a random base number is generated with the GUID as the seed in the extraction implementation stage, so as to minimize the interference of the reproducibility and correlation of pseudo-random numbers.

[0062] 3. Strong applicability: The extraction process is divided into a preprocessing stage and an extraction implementation stage, thus realizing the decoupling of the sample set collation and the extraction logic, ensuring that this extraction method is applicable to various business type sample sets (weighted, unweighted), and various extraction requirements (extracting 1 at a time, extracting multiple ordered samples at a time). Description of the Drawings

[0063] The accompanying drawings described herein are used to provide a further understanding of the embodiments of the present invention, and constitute a part of the present invention, but do not limit the embodiments of the present invention;

[0064] Figure 1 It is a schematic flow chart of a weighted random expert extraction method based on the A-Res algorithm;

[0065] Figure 2 It is a statistical chart of the relationship between sample weights and the number of times drawn;

[0066] Figure 3 It is a schematic architecture diagram of a weighted random expert extraction system based on the A-Res algorithm. Detailed implementation manners

[0067] In order to more clearly understand the above-mentioned objects, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners. It should be noted that, without conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0068] Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0069] Embodiment 1;

[0070] Please refer to Figure 1 , Figure 1 It is a schematic flow chart of a weighted random expert extraction method based on the A-Res algorithm; Embodiment 1 of the present invention provides a weighted random expert extraction method based on the A-Res algorithm, and the method includes:

[0071] Step 1: Collect and store expert information;

[0072] Step 2: Obtain a sample set based on the stored expert information, set extraction conditions, and obtain an expert extraction result from the sample set based on the extraction conditions;

[0073] Step 3: Statistically analyze the expert extraction results obtained by the extraction unit;

[0074] Among them, obtaining an expert extraction result from the sample set based on the extraction conditions specifically includes:

[0075] Obtain a first sample set;

[0076] Randomly arrange the first sample to obtain a second sample set;

[0077] Calculate the total weight of all samples in the second sample set;

[0078] Each sample in the second sample set generates a random base number respectively;

[0079] Based on the total weight sum of all samples, the random base number of each sample, and the sample weight of each sample, calculate the first probability weight value corresponding to each sample respectively;

[0080] Update the first probability weight value of each sample, and each sample obtains the corresponding second probability weight value respectively;

[0081] Based on the second probability weight value of each sample, sort in descending order of the second probability weight value to generate an extraction result sequence;

[0082] Select the first several samples from the extraction result sequence to obtain the expert extraction result.

[0083] Among them, in practical applications, the expert can also be a target such as an agency, a framework agreement manufacturer, and a supplier, etc., which can be adjusted according to actual needs, and the embodiments of the present invention do not make specific limitations.

[0084] Currently, various sample sets have the need for weighted random extraction, such as extraction from a review expert database, extraction from a tender agency, extraction from a framework agreement manufacturer, etc. These sample sets all have the characteristics of low total volume, dynamic change (fixed within a period of time), reflecting the reward and punishment mechanism through weights, strictly ensuring fairness and justice in the extraction process and results, and changing the extraction quantity, etc. Therefore, the present invention designs a weighted random extraction system based on the A-Res algorithm and improves the implementation process, reduces the complexity of code implementation, improves the applicability of the extraction logic, and reduces the influence of the correlation of computer pseudo-random numbers.

[0085] The embodiments of the present invention provide an efficient and fair weighted random extraction system. Through the dual mechanisms of A-Res preprocessing probability weights and random sorting of the sample set, without significantly increasing the algorithm complexity, the code implementation complexity is reduced, and at the same time, the interference of pseudo-random numbers is reduced to ensure the independence and fairness of the sampling results.

[0086] Among them, the core of this method is the extraction method in step 2. The extraction process is introduced in detail below:

[0087] The core idea of step 2 is to combine the sample weight with a random number to generate a probability weight value, and reduce the interference of pseudo-random numbers through random sorting before sampling to achieve efficient and fair weighted random extraction. The specific steps of one extraction are as follows:

[0088] 1. Sample set maintenance:

[0089] Based on the database, relevant business personnel regularly maintain the information of each element in the database (add, delete, modify), including the maintenance of its extraction weight. The database records the key-value pairs of the ID value ( ), and the extraction weight ( ) of each element (i.e., sample), thus forming a complete first sample set :

[0090] ;

[0091] Among them, is the ID value of the m-th sample in the first sample set, 1 ≤ m ≤ n, where n is the total number of samples in the first sample set, is the extraction weight of the m-th sample in the first sample set.

[0092] 2. Sampling preprocessing:

[0093] Random sorting of the sample set: Shuffle the sample set randomly to form a new sample sequence. This is to ensure that the initial order of each extraction is independent, thus reducing the interference of the correlation of pseudo-random numbers in each random extraction.

[0094] Weight accumulation: Calculate the total weight of all samples:

[0095] ;

[0096] Among them, W is the total weight of all samples, is the extraction weight of the i-th sample in the second sample set, 1 ≤ i ≤ n, where n is the total number of samples in the first sample set.

[0097] 3. Sampling implementation:

[0098] Generate a random base number: Use the globally unique identifier (GUID) as the seed (to avoid generating the same random numbers due to the same seed, thus ensuring that each generated random number is independent of each other), and generate a uniform random number between 0 and 1 for each sample (in this method, the random base number is the uniform random number generated by the computer function, that is, the pseudo-random number):

[0099] ;

[0100] Among them, is the random base number of each sample in the second sample set, 1 ≤ i ≤ n, where n is the total number of samples in the first sample set, The function represents taking a random value between 0 and 1.

[0101] Generate the first probability weight: Calculate the first probability weight corresponding to each sample based on the total weight of all samples, the random base number of each sample, and the sample weight of each sample;

[0102] The calculation method of the first probability weight is:

[0103] ;

[0104] Among them, is the first probability weight of the i-th sample in the second sample set, W is the total weight of all samples, is the extraction weight of the i-th sample in the second sample set, is the random base number of each sample in the second sample set, 1 ≤ i ≤ n, and n is the total number of samples in the first sample set.

[0105] Since is between 0 and 1, so when remains unchanged, the larger, the larger the value, and the higher the probability of being selected. However, when the sample size is very large, it will cause the exponent to be very large, resulting in losing precision. Therefore, take the logarithm of the first probability weight to form the second probability weight, which can ensure the precision without affecting the comparison result. The calculation method of the second probability weight is:

[0106] ;

[0107] Among them, is the second probability weight of the i-th sample in the second sample set, W is the total weight of all samples, is the extraction weight of the i-th sample in the second sample set, is the random base number of each sample in the second sample set, 1 ≤ i ≤ n, and n is the total number of samples in the first sample set.

[0108] Sorting and selection: All samples are sorted in descending order according to the generated second probability weight, and the extraction result sequence is R. The expression of R is:

[0109] ;

[0110] Among them, is the ID value of the i-th sample in the second sample set, is the second probability weight of the i-th sample in the second sample set, is the second probability weight of the i - 1-th sample in the second sample set, is the second probability weight of the (i + 1)-th sample in the second sample set, W is the total weight of all samples, 1 ≤ i ≤ n, and n is the total number of samples in the first sample set.

[0111] Take the first k samples of this sequence as the extraction result of this extraction. The value of k can be adjusted according to actual needs, and the embodiments of the present invention do not make specific limitations.

[0112] The sample extraction process in this system is as follows: In the extraction preprocessing stage, the data input is the service sample set and the quantity to be extracted, and the output is the random sequence sample set, the weight accumulation and the quantity to be extracted, which are used as the data input in the extraction implementation stage. The output of the extraction implementation stage is several ordered samples drawn.

[0113] Taking the review expert management system as an example, the system in the present invention is introduced. When extracting review experts, this system directly calls the extraction unit to implement the weighted random extraction of the specified number of review experts.

[0114] Obtain the statistical chart of the relationship between the sample weight and the number of times being drawn;

[0115] From a sample set with n samples and sample weights of each time a sample is randomly drawn, and independently drawn M times, then the theoretical number of times each sample is drawn is:

[0116] ;

[0117] Please refer to Figure 2 Figure 2 is the statistical chart of the relationship between the sample weight and the number of times being drawn;

[0118] Such as Figure 2 statistics show that there are 10 samples in the sample set. Set the weight of a certain sample to change continuously, and the weights of the remaining samples remain unchanged (all are 1). Statistically count the actual number of times this sample is drawn and the theoretical number of times. It can be seen from the statistical chart that by using the weighted random sampling of the present invention, there is a positive correlation between the sample weight and the number of times being drawn, and the actual number of times being drawn is close to the theoretical number of times, indicating that the extraction result is credible and conforms to the probability statistical law.

[0119] The following uses an example of extracting review experts to illustrate the implementation process of the present invention:

[0120] Scenario: Randomly extract 3 review experts from an expert database (assuming 100 experts), and the weight of each expert is based on historical scores (the weight value range is (0, 2], and the default is 1).

[0121] 1. Preprocessing stage;

[0122] Random sorting: Randomly shuffle the expert database to form a random sequence expert database:​

[0123] ;

[0124] Weight accumulation: ;

[0125] 2. Sampling implementation stage;

[0126] Generate probability weights: Based on the extraction weight and random base number of each expert, generate the probability weight of each expert:

[0127] For expert 31 , then ;

[0128] For expert 11 , then ;

[0129] For expert 3 , then ;

[0130]

[0131]

[0132] For expert 64 , then ;

[0133] For expert 28 , then ;

[0134] Sorting and selection: Sort the expert database in descending order of probability weights, and select the top 3 experts, which are the extraction results of this time:

[0135] .

[0136] Example 2;

[0137] Please refer to Figure 3 , Example 2 of the present invention provides a weighted random expert extraction system based on the A-Res algorithm. The system includes:

[0138] A database for collecting expert information, storing expert information, providing expert information retrieval services, and maintaining the stored expert information;

[0139] An extraction unit for obtaining a sample set from the database, setting extraction conditions, and extracting an expert extraction result from the sample set based on the extraction conditions;

[0140] A statistical unit for statistically analyzing the expert extraction result obtained by the extraction unit;

[0141] Among them, the method for obtaining the expert extraction result is as follows:

[0142] Obtain a first sample set;

[0143] Randomly arrange the first samples to obtain a second sample set;

[0144] Calculate the total weight of all samples in the second sample set;

[0145] Each sample in the second sample set generates a random base number respectively;

[0146] Based on the total weight of all samples, the random base number of each sample, and the sample weight of each sample, calculate the first probability weight value corresponding to each sample respectively;

[0147] Update the first probability weight value of each sample, and each sample obtains the corresponding second probability weight value respectively;

[0148] Based on the second probability weight value of each sample, sort them in descending order of the second probability weight value to generate an extraction result sequence;

[0149] Select the first several samples from the extraction result sequence to obtain the expert extraction result.

[0150] Among them, in the embodiment of the present invention, the system further includes a basic configuration unit for performing basic configuration and setting on the system. The basic configuration unit is an existing unit, and corresponding details are not elaborated in the embodiment of the present invention.

[0151] In actual application, the system may further include a display for displaying the expert extraction result, and a communication unit for sending the expert extraction result to a preset communication terminal, etc.

[0152] Among them, the acquisition of the expert extraction result can be implemented by using the Figure 3 extraction module therein.

[0153] The present invention proposes a weighted random extraction method and system based on the A-Res algorithm. Through the dual mechanisms of A-Res preprocessing probability weight values and random sorting of the sample set, without significantly increasing the algorithm complexity, it reduces the code implementation complexity, and at the same time reduces the interference of pseudo-random numbers, ensuring the independence and fairness of the sampling results. It is applicable to efficient and fair random selection scenarios of weighted sample sets with reward and punishment interest relationships such as review experts, agency organizations, and framework agreement manufacturers.

[0154] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn the basic creative concept. Therefore, the appended claims are intended to be interpreted to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present invention.

[0155] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A weighted random expert extraction method based on the A-Res algorithm, characterized in that, The method includes: Step 1: Collect and store expert information; Step 2: Obtain a sample set based on the stored expert information, set extraction conditions, and extract an expert extraction result from the sample set based on the extraction conditions; Step 3: Statistically analyze the expert extraction results obtained by the extraction unit; Among them, extracting the expert extraction result from the sample set based on the extraction conditions specifically includes: Obtain a first sample set; Randomly arrange the first sample to obtain a second sample set; Calculate the total weight sum of all samples in the second sample set; Each sample in the second sample set generates a random base number respectively; Based on the total weight sum of all samples, the random base number of each sample, and the sample weight of each sample, calculate the first probability weight value corresponding to each sample respectively; Update the first probability weight value of each sample, and each sample obtains the corresponding second probability weight value respectively; Based on the second probability weight value of each sample, sort them in descending order of the second probability weight value to generate an extraction result sequence; Select the first several samples from the extraction result sequence to obtain the expert extraction result.

2. The weighted random expert extraction method based on the A-Res algorithm according to claim 1, wherein Step 1 further includes: maintaining the stored expert information, specifically including: Adding, deleting, and changing the stored expert information, and maintaining the extraction weight of the expert information.

3. A weighted random expert extraction method based on the A-Res algorithm according to claim 1, characterized in that, The first sample set is S, and the expression of S is: ； Among them, is the ID value of the m-th sample in the first sample set, where 1 ≤ m ≤ n and n is the total number of samples in the first sample set, is the extraction weight of the m-th sample in the first sample set.

4. A weighted random expert extraction method based on the A-Res algorithm according to claim 3, characterized in that, The calculation method of the total weight sum of all samples is: ； Among them, W is the total weight of all samples, is the extraction weight of the i-th sample in the second sample set, where 1 ≤ i ≤ n and n is the total number of samples in the first sample set.

5. A weighted random expert extraction method based on the A-Res algorithm according to claim 1, characterized in that, The random base number of each sample in the second sample set is generated in the following way: ； wherein, is the random base number for each sample in the second sample set, 1 ≤ i ≤ n, and n is the total number of samples in the first sample set, The function represents taking a random value between 0 and 1.

6. A weighted random expert extraction method based on the A-Res algorithm according to claim 1, characterized in that The calculation method of the first probability weight value is: ； Among them, is the first probability weight of the i-th sample in the second sample set, W is the total weight of all samples, is the sampling weight of the i-th sample in the second sample set, is the random base number of each sample in the second sample set, 1 ≤ i ≤ n, and n is the total number of samples in the first sample set.

7. A weighted random expert extraction method based on the A-Res algorithm according to claim 1, characterized in that The calculation method of the second probability weight value is: ； Among them, is the second probability weight of the i-th sample in the second sample set, W is the total weight of all samples, is the extraction weight of the i-th sample in the second sample set, is the random base number of each sample in the second sample set, 1 ≤ i ≤ n, and n is the total number of samples in the first sample set.

8. A weighted random expert extraction method based on the A-Res algorithm according to claim 1, characterized in that, The extraction result sequence is R, and the expression of R is: ； wherein, is the ID value of the i-th sample in the second sample set, is the second probability weight of the i-th sample in the second sample set, is the second probability weight of the (i - 1)-th sample in the second sample set, is the second probability weight of the (i + 1)-th sample in the second sample set, W is the total weight of all samples, 1 ≤ i ≤ n, and n is the total number of samples in the first sample set.

9. A weighted random expert extraction system based on the A-Res algorithm, characterized in that, The system includes: A database, which is used to collect expert information, store expert information, provide expert information retrieval services, and maintain the stored expert information; An extraction unit, which is used to obtain a sample set from the database, set extraction conditions, and extract an expert extraction result from the sample set based on the extraction conditions; A statistical unit, which is used to statistically analyze the expert extraction results obtained by the extraction unit; Among them, the way to obtain the expert extraction result is: Obtain a first sample set; Randomly arrange the first sample to obtain a second sample set; Calculate the total weight sum of all samples in the second sample set; Each sample in the second sample set generates a random base number respectively; Based on the total weight sum of all samples, the random base number of each sample, and the sample weight of each sample, calculate the first probability weight value corresponding to each sample respectively; Update the first probability weight value of each sample, and each sample obtains the corresponding second probability weight value respectively; Based on the second probability weight value of each sample, sort them in descending order of the second probability weight value to generate an extraction result sequence; Select the first several samples from the extraction result sequence to obtain the expert extraction result.

10. A weighted random expert extraction system based on the A-Res algorithm according to claim 9, characterized in that, The system further includes a basic configuration unit, which is used to perform basic configuration and settings on the system.

Images