Enterprise management consultation data analysis system and method based on machine learning

Abstract

The invention relates to the technical field of data processing, in particular to an enterprise management consultation data analysis system and method based on machine learning. Performing enterprise management consultation content search on the acquired consultation content keyword feature data and the enterprise management feature data through a depth-first search algorithm to generate enterprise management consultation content feature data; and inputting the enterprise management consultation content feature data into a constructed Gaussian naive Bayesian model to carry out enterprise management consultation content data analysis processing, generating enterprise management consultation content analysis result feature data, and pushing the enterprise management consultation content analysis result feature data to a data analysis result display platform through a wireless communication network. Scientific analysis of enterprise management consultation data and intelligent pushing of enterprise management consultation result data are realized.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an enterprise management consulting data analysis system and method based on machine learning. Background Art

[0002] Business management consulting is a professional consulting service that helps companies and entrepreneurs to discover and solve business management and operational problems, improve corporate competitiveness, strengthen learning and implement changes to achieve corporate goals.

[0003] The Chinese invention patent with publication number CN117216192B introduces an enterprise consulting management system, including a consulting module, a keyword extraction module, a matching module and a business processing module; the consulting module is used to provide customers with a dialog box for inputting consulting information, and is used to send the consulting information to the keyword extraction module; the keyword extraction module is used to obtain keywords in the consulting information using a variety of keyword acquisition algorithms with different time complexities; the matching module is used to determine the professional who provides consulting services to the customer based on the keywords in the consulting information, and send a notification message to the business processing module used by the professional; the business processing module is used by the professional to provide consulting services to the customer based on the notification message; the time cost of the enterprise consulting process is high, it is difficult to feedback customer needs in a timely manner, and it is impossible to realize intelligent analysis of enterprise consulting information. Summary of the invention

[0004] 1. Technical issues to be solved

[0005] In order to solve the deficiencies in the background technology, the present invention provides a system and method for analyzing enterprise management consulting data based on machine learning, which can realize scientific analysis of enterprise management consulting data and intelligently push the analysis results of enterprise management consulting data.

[0006] (II) Technical solution

[0007] A method for analyzing enterprise management consulting data based on machine learning, comprising the following steps:

[0008] S1. Collect keyword feature data of consulting content and enterprise management feature data;

[0009] S2, performing enterprise management consulting content search processing based on the consulting content keyword feature data and the enterprise management feature data to generate enterprise management consulting content feature data;

[0010] S3. Collect relevant sample data of enterprise management consulting content feature data to generate enterprise management consulting content sample feature data;

[0011] S4. Constructing a final Gaussian Naive Bayes model based on the sample characteristic data of the enterprise management consulting content;

[0012] S5, inputting the enterprise management consulting content feature data into the constructed final Gaussian Naive Bayes model to perform enterprise management consulting content data analysis and processing, and generating enterprise management consulting content analysis result feature data;

[0013] S6. Construct enterprise management consulting result data based on the consulting content keyword feature data, the enterprise management consulting content feature data and the enterprise management consulting content analysis result feature data, and push it to the data analysis result display platform.

[0014] The present invention receives the user's consultation request and consultation content information through an enterprise management consulting platform and uses the TextRank algorithm to extract keywords in the consultation content information, thereby ensuring the reliability of the extraction result; performs enterprise management consulting content search through a depth-first search algorithm to generate enterprise management consulting content feature data, thereby ensuring the accuracy of the search results; constructs a Gaussian naive Bayes model based on the collected enterprise management consulting content sample feature data, thereby providing a tool for enterprise management consulting data analysis; optimizes the parameters of the Gaussian naive Bayes model through an artificial fish swarm algorithm, thereby improving the speed and efficiency of the optimization process and ensuring the accuracy of the optimization result; inputs the enterprise management consulting content feature data into the constructed Gaussian naive Bayes model to perform enterprise management consulting content data analysis and processing, thereby generating enterprise management consulting content analysis result feature data, thereby realizing scientific analysis of enterprise management consulting data; constructs enterprise management consulting result data, and pushes it to a data analysis result display platform through a wireless communication network, thereby realizing intelligent push of enterprise management consulting result data.

[0015] Preferably, the specific steps of collecting consulting content keyword feature data and enterprise management feature data are as follows:

[0016] S11, receiving the user's consulting request and consulting content information through the enterprise management consulting platform, and after receiving the user's consulting request, the enterprise management consulting platform uses the TextRank algorithm to extract keywords in the consulting content information, and generates a consulting content keyword feature data set A = {a1, a2, ..., a i ,…,a k}, where a i represents the i-th consulting content keyword feature data, k represents the total number of consulting content keyword feature data;

[0017] S12. Generate enterprise management characteristic data B by calling the enterprise management data in the management data storage unit of the relevant enterprise.

[0018] The present invention receives the user's consultation request and consultation content information through the enterprise management consultation platform and uses the TextRank algorithm to extract keywords in the consultation content information to ensure the reliability of the extraction result.

[0019] Preferably, the specific steps of performing enterprise management consulting content search processing based on the consulting content keyword feature data and the enterprise management feature data to generate the enterprise management consulting content feature data are as follows:

[0020] S21, using a depth-first search algorithm to perform enterprise management consulting content search processing on the consulting content keyword feature data in the consulting content keyword feature data set A and the enterprise management feature data B in sequence according to the consulting content keyword number, searching for the enterprise management feature data corresponding to the consulting content keyword feature data in the consulting content keyword feature data set A and generating enterprise management consulting content feature data B through data identification. zixun .

[0021] The enterprise management consulting content is searched through the depth-first search algorithm to generate enterprise management consulting content feature data, ensuring the accuracy of the search results.

[0022] Preferably, the specific steps of collecting relevant sample data of enterprise management consulting content feature data and generating sample feature data of enterprise management consulting content are as follows:

[0023] S31, collecting the enterprise management consulting content feature data B zixun , and construct the enterprise management consulting content sample feature data set C = {c1, c2, ..., c i ,…,c l}, where c i represents the collected characteristic data of the i-th enterprise management consulting content sample, and l represents the total number of collected characteristic data of the enterprise management consulting content sample.

[0024] Preferably, the specific steps of constructing the final Gaussian Naive Bayes model based on the sample characteristic data of the enterprise management consulting content are as follows:

[0025] S41, constructing an initial Gaussian naive Bayes model, setting the number of input nodes of the initial Gaussian naive Bayes model to be a, the number of output nodes to be b, the feature distribution to be Gaussian distribution, and the initial smoothing parameter to be α;

[0026] S42, setting the training data ratio to λ and the test data ratio to η, dividing the enterprise management consulting content sample feature data set according to the training data ratio λ and the test data ratio η, and obtaining an enterprise management consulting content training sample feature data set and an enterprise management consulting content test sample feature data set respectively;

[0027] S43, setting the maximum number of training times to n1 and the training error threshold to n2, inputting the enterprise management consulting content training sample feature data set into the initial Gaussian naive Bayes model for training, until the training error of the initial Gaussian naive Bayes model is less than the training error threshold n2 or the number of training times is greater than the maximum number of training times n1, stopping the training of the initial Gaussian naive Bayes model, and obtaining a trained Gaussian naive Bayes model;

[0028] S44, setting the test accuracy threshold to n3, the first test sample data ratio to h1, and the second test sample data ratio to h2, dividing the enterprise management consulting content test sample feature data set according to the first test sample data ratio h1 and the second test sample data ratio h2, to obtain the first enterprise management consulting content test sample feature data set and the second enterprise management consulting content test sample feature data set respectively;

[0029] Input the first enterprise management consulting content test sample feature data set and the second enterprise management consulting content test sample feature data set into the trained Gaussian naive Bayes model for testing, calculate the accuracy of the test results, and obtain the first test accuracy γ1 and the second test accuracy γ2 respectively;

[0030] When h1·γ1+h2·γ2≥n3, the final Gaussian Naive Bayes model is obtained; otherwise, the parameters of the current Gaussian Naive Bayes model are tuned and the training is returned to S42 for retraining until h1·γ1+h2·γ2≥n3;

[0031] Parameter tuning in S44 includes the following steps:

[0032] S441, construct an artificial fish population, set the artificial fish population size to N, the artificial fish perception range to V, the moving step length to S, the trial number threshold to T, the crowding factor to δ, the current iteration number to t, and the maximum iteration number to t max And the dimension of parameter search space is P;

[0033] Based on the parameters of the current Gaussian Naive Bayes model, N groups of parameters are randomly generated. Each artificial fish in the artificial fish population corresponds to a group of parameters. The initial position set of the artificial fish population in the parameter search space is set to X = {X1, X2, ..., X i ,…,X N}, where X i represents the initial position of the ith artificial fish in the artificial fish population in the parameter search space;

[0034] S442, calculating the fitness function value of each artificial fish in the artificial fish population, arranging the artificial fish in the artificial fish population from large to small according to the fitness function value, and selecting the artificial fish with the highest fitness function value as the current optimal individual; the fitness function formula is as follows:

[0035]

[0036] Among them, F i represents the fitness function value of the ith artificial fish in the artificial fish population, x i represents the ability of the parameters corresponding to the ith artificial fish in the artificial fish population to improve the accuracy of the final Gaussian Naive Bayes model test results, and φ represents the correction value;

[0037] S443, starting an iterative operation, each artificial fish in the artificial fish population will select a flocking behavior or a tail-chasing behavior to update its position;

[0038] S4431, set the number of other artificial fish within the perception range V of the ith artificial fish in the artificial fish population in the current iteration process to n f and the center position is X c ;

[0039] According to the fitness function formula, the fitness function value F of the center position within the perception range V of the i-th artificial fish in the artificial fish population is calculated. c , and the fitness function value of the best individual among other artificial fish within the sensing range V of the ith artificial fish in the artificial fish population is recorded as F best ;

[0040] when And F c ≥F best When , the i-th artificial fish in the artificial fish population will select the clustering behavior in the parameter search space to update the position and move one step to the center position within the perception range; otherwise, no position update is performed; the position update formula is as follows:

[0041]

[0042] in, represents the position of the ith artificial fish in the artificial fish population after the position update in the parameter search space, and rand1 represents a random number between (0,1);

[0043] when And F c <F best , the i-th artificial fish in the artificial fish population will choose the tail-chasing behavior in the parameter search space to update its position and move one step to the position of the best individual among other artificial fish within the sensing range; otherwise, no position update will be performed; the position update formula is as follows:

[0044]

[0045] Among them, X best represents the position of the best individual among other artificial fish within the sensing range V of the ith artificial fish in the artificial fish population, and rand2 represents a random number between (0,1);

[0046] S444, calculating the fitness function value of the ith artificial fish in the artificial fish population after selecting the clustering behavior or the tail-chasing behavior to update the position, if the fitness function value after the position update is greater than the original fitness function value, proceeding to S445; otherwise, the ith artificial fish in the artificial fish population randomly selects a direction in the parameter search space to update the position, until the fitness function value after the position update of the ith artificial fish in the artificial fish population is greater than the original fitness function value or the number of times the ith artificial fish in the artificial fish population randomly selects a direction in the parameter search space to update the position is greater than the trial number threshold T;

[0047] S445, repeating the operations from S443 to S444 until all artificial fish in the artificial fish population are traversed, and the fitness function values ​​of the artificial fish in the artificial fish population after the positions are updated by selecting the clustering behavior or the tail-chasing behavior are arranged from large to small, and the artificial fish with the highest fitness function value is selected as the current optimal individual;

[0048] S446: Determine whether the current number of iterations t is less than the maximum number of iterations t max , if the current number of iterations t is less than the maximum number of iterations t max , then the current iteration number t is increased by 1, and the process returns to S443; otherwise, the current optimal individual is taken as the global optimal solution, the parameters corresponding to the global optimal solution are output as the optimal parameters, and the parameters of the current Gaussian Naive Bayes model are adjusted to the optimal parameters.

[0049] A Gaussian Naive Bayes model is constructed based on the collected characteristic data of enterprise management consulting content samples, which provides a tool for enterprise management consulting data analysis. The parameters of the Gaussian Naive Bayes model are tuned through the artificial fish swarm algorithm. Compared with the traditional parameter search method, the speed and efficiency of the optimization process are improved, the accuracy of the optimization results is ensured, and the foraging behavior of artificial fish is simulated to effectively avoid the optimization results from falling into the local optimal solution.

[0050] Preferably, the enterprise management consulting content feature data is input into the final constructed Gaussian Naive Bayes model to perform enterprise management consulting content data analysis and processing, and the specific steps of generating enterprise management consulting content analysis result feature data are as follows:

[0051] S51, the enterprise management consulting content feature data B zixunInput the final Gaussian Naive Bayes model to analyze and process the enterprise management consulting content data, and generate the characteristic data D of the enterprise management consulting content analysis results fenxi .

[0052] The enterprise management consulting content feature data is input into the constructed Gaussian Naive Bayes model to perform enterprise management consulting content data analysis and processing, generate enterprise management consulting content analysis result feature data, and realize scientific analysis of enterprise management consulting data.

[0053] Preferably, the specific steps of constructing enterprise management consulting result data based on the consulting content keyword feature data, the enterprise management consulting content feature data and the enterprise management consulting content analysis result feature data, and pushing it to the data analysis result display platform are as follows:

[0054] S61, the consulting content keyword feature data set A, the enterprise management consulting content feature data set B zixun And the characteristic data D of the enterprise management consulting content analysis result fenxi Combine the data to construct the enterprise management consulting result data Q = (A, B zixun ,D fenxi );

[0055] S62, transmitting the enterprise management consulting result data Q=(A, B zixun ,D fenxi ) is pushed to the data analysis result display platform.

[0056] Build enterprise management consulting result data and push it to the data analysis result display platform through the wireless communication network to realize the intelligent push of enterprise management consulting result data.

[0057] The present invention also discloses a business management consulting data analysis system based on machine learning, including a data acquisition module, a business management consulting content feature data search module, a business management consulting content sample feature data collection module, a final Gaussian naive Bayes model construction module, a parameter optimization module, a data analysis module and a business management consulting result data output module;

[0058] After receiving the user's consultation request through the enterprise management consultation platform, the data collection module uses the TextRank algorithm to extract keywords in the consultation content information, generates consultation content keyword feature data, and calls the enterprise management data in the management data storage unit of the relevant enterprise to generate enterprise management feature data;

[0059] The enterprise management consulting content feature data search module performs enterprise management consulting content search processing on the consulting content keyword feature data and the enterprise management feature data through a depth-first search algorithm to generate enterprise management consulting content feature data;

[0060] The enterprise management consulting content sample feature data collection module constructs an enterprise management consulting content sample feature data set by collecting relevant data of the enterprise management consulting content feature data;

[0061] The final Gaussian Naive Bayesian model construction module constructs a final Gaussian Naive Bayesian model based on the enterprise management consulting content sample feature data set;

[0062] The parameter optimization module optimizes the parameters of the current Gaussian Naive Bayes model through an artificial fish swarm optimization algorithm;

[0063] The data analysis module inputs the enterprise management consulting content feature data into the final Gaussian Naive Bayes model to perform enterprise management consulting content data analysis and processing, and generates enterprise management consulting content analysis result feature data;

[0064] The enterprise management consulting result data output module combines the consulting content keyword feature data, the enterprise management consulting content feature data and the enterprise management consulting content analysis result feature data to construct the enterprise management consulting result data, and pushes it to the data analysis result display platform through the wireless communication network.

[0065] (III) Beneficial effects

[0066] 1. The present invention sets a data acquisition module, an enterprise management consulting content feature data search module, an enterprise management consulting content sample feature data collection module, a final Gaussian naive Bayes model construction module, a parameter optimization module, a data analysis module and an enterprise management consulting result data output module; receives the user's consulting request and consulting content information through the enterprise management consulting platform and uses the TextRank algorithm to extract the keywords in the consulting content information to ensure the reliability of the extraction results; searches for enterprise management consulting content through a depth-first search algorithm to generate enterprise management consulting content feature data, thereby ensuring the accuracy of the search results; constructs a Gaussian naive Bayes model based on the collected enterprise management consulting content sample feature data, thereby providing a tool for enterprise management consulting data analysis; inputs the enterprise management consulting content feature data into the constructed Gaussian naive Bayes model to perform enterprise management consulting content data analysis and processing, generates enterprise management consulting content analysis result feature data, and realizes scientific analysis of enterprise management consulting data; constructs enterprise management consulting result data, and pushes it to the data analysis result display platform through a wireless communication network, thereby realizing intelligent push of enterprise management consulting result data;

[0067] 2. The artificial fish swarm algorithm is used to optimize the parameters of the Gaussian Naive Bayes model. Compared with the traditional parameter search method, the speed and efficiency of the optimization process are improved, the accuracy of the optimization results is ensured, and the foraging behavior of artificial fish is simulated to effectively avoid the optimization results from falling into the local optimal solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0069] Figure 1 A flowchart of a method for analyzing enterprise management consulting data based on machine learning provided by the present invention;

[0070] Figure 2 A module schematic diagram of a business management consulting data analysis system based on machine learning provided by the present invention. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0073] Embodiment 1 is as follows:

[0074] See also Figure 1 , a method for analyzing enterprise management consulting data based on machine learning, comprising the following steps:

[0075] S1. Collect keyword feature data of consulting content and enterprise management feature data;

[0076] S11, receiving the user's consulting request and consulting content information through the enterprise management consulting platform, and after receiving the user's consulting request, the enterprise management consulting platform uses the TextRank algorithm to extract keywords in the consulting content information, and generates a consulting content keyword feature data set A = {a1, a2, ..., a i ,…,a k}, where a i represents the i-th consulting content keyword feature data, k represents the total number of consulting content keyword feature data;

[0077] S12. Generate enterprise management characteristic data B by calling the enterprise management data in the management data storage unit of the relevant enterprise.

[0078] S2, performing enterprise management consulting content search processing based on the consulting content keyword feature data and the enterprise management feature data to generate enterprise management consulting content feature data;

[0079] S21, using a depth-first search algorithm to perform enterprise management consulting content search processing on the consulting content keyword feature data in the consulting content keyword feature data set A and the enterprise management feature data B in sequence according to the consulting content keyword number, searching for the enterprise management feature data corresponding to the consulting content keyword feature data in the consulting content keyword feature data set A and generating enterprise management consulting content feature data B through data identification. zixun .

[0080] S3. Collect relevant sample data of enterprise management consulting content feature data to generate enterprise management consulting content sample feature data;

[0081] S31, collecting the enterprise management consulting content feature data B zixun , and construct the enterprise management consulting content sample feature data set C = {c1, c2, ..., c i ,…,c l}, where c i represents the collected characteristic data of the i-th enterprise management consulting content sample, and l represents the total number of collected characteristic data of the enterprise management consulting content sample.

[0082] S4. Constructing a final Gaussian Naive Bayes model based on the sample characteristic data of the enterprise management consulting content;

[0083] S41, constructing an initial Gaussian naive Bayes model, setting the number of input nodes of the initial Gaussian naive Bayes model to be a, the number of output nodes to be b, the feature distribution to be Gaussian distribution, and the initial smoothing parameter to be α;

[0084] S42, setting the training data ratio to λ and the test data ratio to η, dividing the enterprise management consulting content sample feature data set according to the training data ratio λ and the test data ratio η, and obtaining an enterprise management consulting content training sample feature data set and an enterprise management consulting content test sample feature data set respectively;

[0085] S43, setting the maximum number of training times to n1 and the training error threshold to n2, inputting the enterprise management consulting content training sample feature data set into the initial Gaussian naive Bayes model for training, until the training error of the initial Gaussian naive Bayes model is less than the training error threshold n2 or the number of training times is greater than the maximum number of training times n1, stopping the training of the initial Gaussian naive Bayes model, and obtaining a trained Gaussian naive Bayes model;

[0086] S44, setting the test accuracy threshold to n3, the first test sample data ratio to h1, and the second test sample data ratio to h2, dividing the enterprise management consulting content test sample feature data set according to the first test sample data ratio h1 and the second test sample data ratio h2, to obtain the first enterprise management consulting content test sample feature data set and the second enterprise management consulting content test sample feature data set respectively;

[0087] Input the first enterprise management consulting content test sample feature data set and the second enterprise management consulting content test sample feature data set into the trained Gaussian naive Bayes model for testing, calculate the accuracy of the test results, and obtain the first test accuracy γ1 and the second test accuracy γ2 respectively;

[0088] When h1·γ1+h2·γ2≥n3, the final Gaussian Naive Bayes model is obtained; otherwise, the parameters of the current Gaussian Naive Bayes model are tuned and the training is returned to S42 for retraining until h1·γ1+h2·γ2≥n3;

[0089] Parameter tuning in S44 includes the following steps:

[0090] S441, construct an artificial fish population, set the artificial fish population size to N, the artificial fish perception range to V, the moving step length to S, the trial number threshold to T, the crowding factor to δ, the current iteration number to t, and the maximum iteration number to t max And the dimension of parameter search space is P;

[0091] Based on the parameters of the current Gaussian Naive Bayes model, N groups of parameters are randomly generated. Each artificial fish in the artificial fish population corresponds to a group of parameters. The initial position set of the artificial fish population in the parameter search space is set to X = {X1, X2, ..., X i ,…,X N}, where Xi represents the initial position of the ith artificial fish in the artificial fish population in the parameter search space;

[0092] S442, calculating the fitness function value of each artificial fish in the artificial fish population, arranging the artificial fish in the artificial fish population from large to small according to the fitness function value, and selecting the artificial fish with the highest fitness function value as the current optimal individual; the fitness function formula is as follows:

[0093]

[0094] Among them, F i represents the fitness function value of the ith artificial fish in the artificial fish population, x i represents the ability of the parameters corresponding to the ith artificial fish in the artificial fish population to improve the accuracy of the final Gaussian Naive Bayes model test results, and φ represents the correction value;

[0095] S443, starting an iterative operation, each artificial fish in the artificial fish population will select a flocking behavior or a tail-chasing behavior to update its position;

[0096] S4431, set the number of other artificial fish within the perception range V of the ith artificial fish in the artificial fish population in the current iteration process to n f and the center position is X c ;

[0097] According to the fitness function formula, the fitness function value F of the center position within the perception range V of the i-th artificial fish in the artificial fish population is calculated. c , and the fitness function value of the best individual among other artificial fish within the sensing range V of the ith artificial fish in the artificial fish population is recorded as F best ;

[0098] when And F c ≥F best When , the i-th artificial fish in the artificial fish population will select the clustering behavior in the parameter search space to update the position and move one step to the center position within the perception range; otherwise, no position update is performed; the position update formula is as follows:

[0099]

[0100] in, represents the position of the ith artificial fish in the artificial fish population after the position update in the parameter search space, and rand1 represents a random number between (0,1);

[0101] when And F c <F best, the i-th artificial fish in the artificial fish population will choose the tail-chasing behavior in the parameter search space to update its position and move one step to the position of the best individual among other artificial fish within the sensing range; otherwise, no position update will be performed; the position update formula is as follows:

[0102]

[0103] Among them, X best represents the position of the best individual among other artificial fish within the sensing range V of the ith artificial fish in the artificial fish population, and rand2 represents a random number between (0,1);

[0104] S444, calculating the fitness function value of the ith artificial fish in the artificial fish population after selecting the clustering behavior or the tail-chasing behavior to update the position, if the fitness function value after the position update is greater than the original fitness function value, proceeding to S445; otherwise, the ith artificial fish in the artificial fish population randomly selects a direction in the parameter search space to update the position, until the fitness function value after the position update of the ith artificial fish in the artificial fish population is greater than the original fitness function value or the number of times the ith artificial fish in the artificial fish population randomly selects a direction in the parameter search space to update the position is greater than the trial number threshold T;

[0105] S445, repeating the operations from S443 to S444 until all artificial fish in the artificial fish population are traversed, and the fitness function values ​​of the artificial fish in the artificial fish population after the positions are updated by selecting the clustering behavior or the tail-chasing behavior are arranged from large to small, and the artificial fish with the highest fitness function value is selected as the current optimal individual;

[0106] S446: Determine whether the current number of iterations t is less than the maximum number of iterations t max , if the current number of iterations t is less than the maximum number of iterations t max , then the current iteration number t is increased by 1, and the process returns to S443; otherwise, the current optimal individual is taken as the global optimal solution, the parameters corresponding to the global optimal solution are output as the optimal parameters, and the parameters of the current Gaussian Naive Bayes model are adjusted to the optimal parameters.

[0107] S5, inputting the enterprise management consulting content feature data into the constructed final Gaussian Naive Bayes model to perform enterprise management consulting content data analysis and processing, and generating enterprise management consulting content analysis result feature data;

[0108] S51, the enterprise management consulting content feature data B zixun Input the final Gaussian Naive Bayes model to analyze and process the enterprise management consulting content data, and generate the characteristic data D of the enterprise management consulting content analysis results fenxi .

[0109] S6. Constructing enterprise management consulting result data based on the consulting content keyword feature data, the enterprise management consulting content feature data and the enterprise management consulting content analysis result feature data, and pushing it to the data analysis result display platform;

[0110] S61, the consulting content keyword feature data set A, the enterprise management consulting content feature data set B zixun And the characteristic data D of the enterprise management consulting content analysis result fenxi Combine the data to construct the enterprise management consulting result data Q = (A, B zixun ,D fenxi );

[0111] S62, transmitting the enterprise management consulting result data Q=(A, B zixun ,D fenxi ) is pushed to the data analysis result display platform.

[0112] Embodiment 2 is as follows:

[0113] See also Figure 2 , a business management consulting data analysis system based on machine learning, including a data acquisition module, a business management consulting content feature data search module, a business management consulting content sample feature data collection module, a final Gaussian naive Bayes model construction module, a parameter optimization module, a data analysis module and a business management consulting result data output module;

[0114] After receiving the user's consultation request through the enterprise management consultation platform, the data collection module uses the TextRank algorithm to extract keywords in the consultation content information, generates consultation content keyword feature data, and calls the enterprise management data in the management data storage unit of the relevant enterprise to generate enterprise management feature data;

[0115] The enterprise management consulting content feature data search module performs enterprise management consulting content search processing on the consulting content keyword feature data and the enterprise management feature data through a depth-first search algorithm to generate enterprise management consulting content feature data;

[0116] The enterprise management consulting content sample feature data collection module constructs an enterprise management consulting content sample feature data set by collecting relevant data of the enterprise management consulting content feature data;

[0117] The final Gaussian Naive Bayesian model construction module constructs a final Gaussian Naive Bayesian model based on the enterprise management consulting content sample feature data set;

[0118] The parameter optimization module optimizes the parameters of the current Gaussian Naive Bayes model through an artificial fish swarm optimization algorithm;

[0119] The data analysis module inputs the enterprise management consulting content feature data into the final Gaussian Naive Bayes model to perform enterprise management consulting content data analysis and processing, and generates enterprise management consulting content analysis result feature data;

[0120] The enterprise management consulting result data output module combines the consulting content keyword feature data, the enterprise management consulting content feature data and the enterprise management consulting content analysis result feature data to construct the enterprise management consulting result data, and pushes it to the data analysis result display platform through the wireless communication network.

[0121] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0122] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. A method for analyzing enterprise management consulting data based on machine learning, characterized in that: The steps include: S1. Collect keyword feature data of consulting content and enterprise management feature data; S2, performing enterprise management consulting content search processing based on the consulting content keyword feature data and the enterprise management feature data to generate enterprise management consulting content feature data; S3. Collect relevant sample data of enterprise management consulting content feature data to generate enterprise management consulting content sample feature data; S4. Constructing a final Gaussian Naive Bayes model based on the sample characteristic data of the enterprise management consulting content; S5, inputting the enterprise management consulting content feature data into the constructed final Gaussian Naive Bayes model to perform enterprise management consulting content data analysis and processing, and generating enterprise management consulting content analysis result feature data; S6. Construct enterprise management consulting result data based on the consulting content keyword feature data, the enterprise management consulting content feature data and the enterprise management consulting content analysis result feature data, and push it to the data analysis result display platform.

2. The enterprise management consulting data analysis method based on machine learning according to claim 1 is characterized in that: The S1 comprises the following steps: S11, receiving the user's consulting request and consulting content information through the enterprise management consulting platform, and after receiving the user's consulting request, the enterprise management consulting platform uses the TextRank algorithm to extract keywords in the consulting content information, and generates a consulting content keyword feature data set A = {a1, a2, ..., a i ,…,a k }, where a i represents the i-th consulting content keyword feature data, k represents the total number of consulting content keyword feature data; S12. Generate enterprise management characteristic data B by calling the enterprise management data in the management data storage unit of the relevant enterprise.

3. The enterprise management consulting data analysis method based on machine learning according to claim 2 is characterized in that: The S2 comprises the following steps: S21, using a depth-first search algorithm to perform enterprise management consulting content search processing on the consulting content keyword feature data in the consulting content keyword feature data set A and the enterprise management feature data B in sequence according to the consulting content keyword number, searching for the enterprise management feature data corresponding to the consulting content keyword feature data in the consulting content keyword feature data set A and generating enterprise management consulting content feature data B through data identification. zixun .

4. The enterprise management consulting data analysis method based on machine learning according to claim 3 is characterized in that: The S3 comprises the following steps: S31, collecting the enterprise management consulting content feature data B zixun , and construct the enterprise management consulting content sample feature data set C = {c1, c2, ..., c i ,…,c l }, where c i represents the collected characteristic data of the i-th enterprise management consulting content sample, and l represents the total number of collected characteristic data of the enterprise management consulting content sample.

5. The enterprise management consulting data analysis method based on machine learning according to claim 4 is characterized in that: The S4 comprises the following steps: S41, constructing an initial Gaussian naive Bayes model, setting the number of input nodes of the initial Gaussian naive Bayes model to be a, the number of output nodes to be b, the feature distribution to be Gaussian distribution, and the initial smoothing parameter to be α; S42, setting the training data ratio to λ and the test data ratio to η, dividing the enterprise management consulting content sample feature data set according to the training data ratio λ and the test data ratio η, and obtaining an enterprise management consulting content training sample feature data set and an enterprise management consulting content test sample feature data set respectively; S43, setting the maximum number of training times to n1 and the training error threshold to n2, inputting the enterprise management consulting content training sample feature data set into the initial Gaussian naive Bayes model for training, until the training error of the initial Gaussian naive Bayes model is less than the training error threshold n2 or the number of training times is greater than the maximum number of training times n1, stopping the training of the initial Gaussian naive Bayes model, and obtaining a trained Gaussian naive Bayes model; S44, setting the test accuracy threshold to n3, the first test sample data ratio to h1, and the second test sample data ratio to h2, dividing the enterprise management consulting content test sample feature data set according to the first test sample data ratio h1 and the second test sample data ratio h2, to obtain the first enterprise management consulting content test sample feature data set and the second enterprise management consulting content test sample feature data set respectively; Input the first enterprise management consulting content test sample feature data set and the second enterprise management consulting content test sample feature data set into the trained Gaussian naive Bayes model for testing, calculate the accuracy of the test results, and obtain the first test accuracy γ1 and the second test accuracy γ2 respectively; When h1·γ1+h2·γ2≥n3, the final Gaussian Naive Bayes model is obtained; otherwise, the parameters of the current Gaussian Naive Bayes model are tuned and returned to S42 for retraining until h1·γ1+h2·γ2≥n3.

6. The enterprise management consulting data analysis method based on machine learning according to claim 5 is characterized in that: The parameter tuning in S44 includes the following steps: S441, construct an artificial fish population, set the artificial fish population size to N, the artificial fish perception range to V, the moving step length to S, the trial number threshold to T, the crowding factor to δ, the current iteration number to t, and the maximum iteration number to t max And the dimension of parameter search space is P; Based on the parameters of the current Gaussian Naive Bayes model, N groups of parameters are randomly generated. Each artificial fish in the artificial fish population corresponds to a group of parameters. The initial position set of the artificial fish population in the parameter search space is set to X = {X1, X2, ..., X i ,…,X N }, where X i represents the initial position of the ith artificial fish in the artificial fish population in the parameter search space; S442, calculating the fitness function value of each artificial fish in the artificial fish population, arranging the artificial fish in the artificial fish population from large to small according to the fitness function value, and selecting the artificial fish with the highest fitness function value as the current optimal individual; S443, starting an iterative operation, each artificial fish in the artificial fish population will select a flocking behavior or a tail-chasing behavior to update its position; S444, calculating the fitness function value of the ith artificial fish in the artificial fish population after selecting the clustering behavior or the tail-chasing behavior to update the position, if the fitness function value after the position update is greater than the original fitness function value, proceeding to S445; otherwise, the ith artificial fish in the artificial fish population randomly selects a direction in the parameter search space to update the position, until the fitness function value after the position update of the ith artificial fish in the artificial fish population is greater than the original fitness function value or the number of times the ith artificial fish in the artificial fish population randomly selects a direction in the parameter search space to update the position is greater than the trial number threshold T; S445, repeating the operations from S443 to S444 until all artificial fish in the artificial fish population are traversed, and the fitness function values ​​of the artificial fish in the artificial fish population after the positions are updated by selecting the clustering behavior or the tail-chasing behavior are arranged from large to small, and the artificial fish with the highest fitness function value is selected as the current optimal individual; S446: Determine whether the current number of iterations t is less than the maximum number of iterations t max , if the current number of iterations t is less than the maximum number of iterations t max , then the current iteration number t is increased by 1, and the process returns to S443; otherwise, the current optimal individual is taken as the global optimal solution, the parameters corresponding to the global optimal solution are output as the optimal parameters, and the parameters of the current Gaussian Naive Bayes model are adjusted to the optimal parameters.

7. The enterprise management consulting data analysis method based on machine learning according to claim 6 is characterized in that: The S443 comprises the following steps: S4431, set the number of other artificial fish within the perception range V of the ith artificial fish in the artificial fish population in the current iteration process to n f and the center position is X c ; According to the fitness function formula, the fitness function value F of the center position within the perception range V of the i-th artificial fish in the artificial fish population is calculated c , and the fitness function value of the best individual among other artificial fish within the sensing range V of the ith artificial fish in the artificial fish population is recorded as F best , the fitness function value of the ith artificial fish in the artificial fish population is recorded as F i ; when And F c ≥F best When , the i-th artificial fish in the artificial fish population will select the clustering behavior in the parameter search space to update its position and move one step toward the center position within the perception range; otherwise, no position update will be performed; when And F c <F best , the i-th artificial fish in the artificial fish population will choose the tail-chasing behavior in the parameter search space to update its position and move one step toward the position of the best individual among other artificial fish within the perception range; otherwise, no position update will be performed.

8. The enterprise management consulting data analysis method based on machine learning according to claim 7 is characterized in that: The S5 comprises the following steps: S51, the enterprise management consulting content feature data B zixun Input the final Gaussian Naive Bayes model to analyze and process the enterprise management consulting content data, and generate the characteristic data D of the enterprise management consulting content analysis results fenxi .

9. The enterprise management consulting data analysis method based on machine learning according to claim 8 is characterized in that: The S6 comprises the following steps: S61, the consulting content keyword feature data set A, the enterprise management consulting content feature data set B zixun And the characteristic data D of the enterprise management consulting content analysis result fenxi Combine the data to construct the enterprise management consulting result data Q = (A, B zixun ,D fenxi ); S62, transmitting the enterprise management consulting result data Q=(A, B zixun ,D fenxi ) is pushed to the data analysis result display platform.

10. A system for implementing the enterprise management consulting data analysis method based on machine learning as described in any one of claims 1 to 9, characterized in that: It includes a data collection module, an enterprise management consulting content feature data search module, an enterprise management consulting content sample feature data collection module, a final Gaussian naive Bayes model construction module, a parameter optimization module, a data analysis module and an enterprise management consulting result data output module.

Images