Enterprise management system based on artificial intelligence technology

Abstract

The invention relates to the technical field of enterprise management, in particular to an enterprise management system based on an artificial intelligence technology, which comprises an intelligent management responsibility sorting module, an intelligent feedback information positioning module and a hierarchical association feedback module, the daily texts with the similarity are classified into one class, a clustering result is expressed through a tree diagram, corresponding administrators A and B are matched according to responsibility sets by analyzing the responsibility sets in the leaf nodes, the administrator A is an administrator in one leaf node, and the administrator B is an administrator corresponding to multiple leaf nodes; the employees are helped to understand the responsibility ranges of different administrators, feedback is accurately pushed, and an enterprise can further improve the organization structure according to clear responsibility division; and the feedback information intelligent positioning module matches the responsibility vocabulary in the feedback text with the responsibility set of each leaf node administrator, and determines the administrator corresponding to the feedback text.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise management, and more specifically, to an enterprise management system based on artificial intelligence technology. Background Art

[0002] With the digital wave sweeping through, enterprise management systems have penetrated into all key aspects of enterprise operations in an all-round way. Among them, artificial intelligence technology is applied and integrated into enterprise management systems, and has a feedback mechanism. Employees can upload feedback texts such as various actual situations encountered in the work process and constructive suggestions to the corresponding feedback database, so that administrators can process and view the feedback texts in the feedback database according to their permissions.

[0003] However, when employees use the enterprise management system, they need to submit feedback texts to relevant administrators (that is, the staff responsible for processing the content of the feedback text. For example, the administrator responsible for salary affairs belongs to the finance department). However, the enterprise management system is not convenient to automatically identify the content of the feedback text for automatic delivery. As a result, when distributing, it can only rely on employees or administrators to search by themselves, which affects the efficiency of enterprise management. In view of this, we propose an enterprise management system based on artificial intelligence technology. Summary of the Invention

[0004] The purpose of the present invention is to solve the problem that in the digital age, enterprise management systems are fully integrated into the key links of enterprise operations, artificial intelligence technology is deeply integrated, significantly improving the system efficiency. Its feedback mechanism builds a communication bridge for enterprises. Employees can organize the actual work situations and suggestions into feedback texts and upload them to the feedback database for administrators to analyze production management problems. At the same time, to improve the work efficiency of administrators, enterprises build a daily database to store their daily processing files. However, due to the complex organizational structure and large number of employees in large enterprises, the business scopes and management focuses of administrators at different levels vary greatly. Although the enterprise has introduced the responsibilities of administrators, due to the complexity of information and poor communication, employees still have difficulty in clarifying specific responsibilities, resulting in the inability to accurately deliver feedback texts to the corresponding administrators, and then causing the problem that feedback texts cannot be concerned and processed for a long time.

[0005] To achieve the above object, the present invention provides an enterprise management system based on artificial intelligence technology, including an intelligent management responsibility sorting module, an intelligent feedback information positioning module, and a hierarchical associated feedback module, wherein:

[0006] The intelligent management responsibility sorting module performs clustering analysis on multiple daily texts, classifies daily texts with similarity into one category, represents the clustering results using a tree diagram, analyzes the responsibility set of each leaf node in the tree diagram, determines the administrator corresponding to the leaf node, and establishes the association relationship between each administrator and the responsibility set; the feedback information intelligent positioning module senses the responsibility vocabulary in the feedback text, matches the responsibility set of the leaf node corresponding to the responsibility vocabulary in the feedback text, and after matching, calls out the administrator corresponding to the leaf node to push the feedback text; the hierarchical association feedback module determines the superior and subordinate of the administrator according to the tree diagram. When pushing the feedback text to the superior administrator, it calls out the subordinate administrator corresponding to the feedback text and pushes it synchronously.

[0007] As a further improvement of this technical solution, the working steps of the clustering analysis in the intelligent management responsibility sorting module are as follows:

[0008] Step 1: Sense the responsibility domain vocabulary in the enterprise organizational structure, and construct a vocabulary dictionary with the responsibility domain vocabulary; traverse each daily text, compare each responsibility domain vocabulary in the vocabulary dictionary one by one. If a responsibility domain vocabulary appears in the daily text, then increase the number of occurrences of the responsibility domain vocabulary in the daily text by ; and according to the order of the responsibility domain vocabulary in the vocabulary dictionary, take the number of occurrences of each responsibility domain vocabulary in the daily text as an element in the daily text vector, and construct the daily text vector corresponding to each daily text.

[0009] Step 2: The clustering analysis is specifically a bottom-up clustering method. Starting from each daily text vector as a separate class, continuously merge similar classes until a class containing all daily text vectors is formed; starting from each daily text vector as a separate cluster, there are clusters, each cluster contains one daily text vector, The quantity of which is equal to the daily text vector data; for each pair of clusters, calculate the similarity between each cluster pair to obtain a similarity matrix, and the similarity matrix records the similarity degree between each two cluster pairs.

[0010] Set a similarity threshold, merge the two clusters in the similarity matrix whose similarity > similarity threshold, and merge the text vectors to form a more representative cluster. After merging, a new cluster is formed, and the new cluster contains all the text vectors in the original two clusters.

[0011] Repeat the above steps of calculating similarity and merging clusters. Each iteration will reduce one cluster until there is no cluster in the similarity matrix with similarity > similarity threshold.

[0012] As a further improvement of this technical solution, the intelligent management responsibility sorting module calculates the similarity between each cluster: It senses that two daily text vectors are respectively and , where is the total number of responsibility area words in the vocabulary dictionary, and are respectively the number of occurrences of the th responsibility area word in the vocabulary dictionary in the daily text and the daily text . The similarity between the daily text is calculated by the following formula: .

[0013] .

[0014] As a further improvement of this technical solution, the leaf nodes of the tree diagram in the intelligent management responsibility sorting module are daily text vectors; each time two or more similar clusters are merged into a new cluster according to the similarity threshold, the new cluster is represented by a non-leaf node in the tree diagram, and each non-leaf node records a cluster merging operation; each non-leaf node represents a cluster merging operation, the branches of the tree diagram represent the connection between different nodes, and the length of the branches represents the similarity between the two clusters during merging.

[0015] As a further improvement of this technical solution, the intelligent management responsibility sorting module retrieves the daily text corresponding to each leaf node in the tree diagram, each responsibility area word in each daily text, constructs a responsibility area word set from the responsibility area words in all leaf nodes, calculates the occurrence frequency of different responsibility area words in the responsibility area word set, removes duplicates from the responsibility area word set according to the occurrence frequency of each responsibility area word. When the same word has different occurrence frequencies in different leaf nodes, the word with the maximum occurrence frequency is saved to the corresponding leaf node, and the de-duplicated responsibility area word set is used as the responsibility set of the leaf node.

[0016] As a further improvement of this technical solution, the working principle of the intelligent management responsibility sorting module for establishing the association relationship between each administrator and the responsibility set is as follows: It senses the daily text corresponding to each leaf node and each administrator to which the daily text belongs, retrieves the administrators whose all daily texts are only in one leaf node, and defines them as administrator A, and associates the responsibility set of the leaf node with administrator A; defines the administrators in multiple leaf nodes as administrator B; and associates the responsibility sets of multiple leaf nodes with administrator B.

[0017] As a further improvement of this technical solution, the feedback information intelligent positioning module adopts the method of extracting key responsibility area vocabulary in daily texts in the management responsibility intelligent sorting module to extract responsibility vocabulary in the feedback text, and compares the responsibility vocabulary in the feedback text with the responsibility sets of each leaf node respectively to match the leaf nodes corresponding to the responsibility vocabulary in the feedback text:

[0018] If the responsibility vocabulary completely matches the responsibility set of a certain leaf node, then call out the administrator A corresponding to the leaf node and output the feedback text to administrator A; if the responsibility vocabulary matches the responsibility sets of multiple leaf nodes respectively, then call out the administrator B corresponding to the leaf node and output the feedback text to administrator B.

[0019] As a further improvement of this technical solution, the feedback information intelligent positioning module senses the responsibility sets of leaf nodes as , where represents the th responsibility vocabulary in the responsibility set. For the responsibility vocabulary set of the feedback text and the responsibility set of the leaf node , if , then find the administrator A corresponding to the leaf node and output the feedback text to administrator A; if there are multiple leaf nodes such that , then find the administrator B corresponding to the leaf node and output the feedback text to administrator B.

[0020] As a further improvement of this technical solution, the hierarchical association feedback module senses the administrator corresponding to each leaf node. If both administrator A and administrator B are included in the leaf nodes, then determine that administrator B is the superior of administrator A.

[0021] As a further improvement of this technical solution, the hierarchical association feedback module senses multiple leaf nodes where the responsibility vocabulary matches, and calls out the administrator A of each leaf node respectively. When the feedback text is output to administrator B, the administrator A and the feedback text are output to administrator B synchronously.

[0022] Compared with the prior art, the beneficial effects of the present invention are:

[0023] In the enterprise management system based on artificial intelligence technology, the intelligent management responsibility sorting module conducts cluster analysis on the daily files of administrators, classifies daily texts with similarity into one category, and represents the clustering results through a tree diagram. According to the responsibility sets within the analysis leaf nodes, corresponding administrators A and B are matched according to the responsibility sets. Administrator A is the administrator in one leaf node, and administrator B is the administrator corresponding to multiple leaf nodes, which helps employees understand the scope of responsibilities of different administrators, accurately push feedback, and also provides data support for the enterprise to optimize the management process, enabling the enterprise to further improve the organizational structure according to clear responsibility divisions and improve management efficiency;

[0024] The intelligent feedback information positioning module matches the responsibility vocabulary in the feedback text with the responsibility sets of each leaf node administrator to determine the administrator corresponding to the feedback text, avoiding the disorderly transfer of feedback information and the situation where administrators are overwhelmed by a large amount of information, improving the feedback processing efficiency and saving time costs. In this process, when the responsibility vocabulary in the feedback text matches the management corresponding to multiple leaf nodes, when the hierarchical association feedback module matches the administrator based on the intelligent management responsibility sorting module, the administrator in multiple leaf nodes within the responsibility set is defined as administrator B. Therefore, when the feedback text is output to administrator B, the corresponding administrator A is output to administrator B at the same time. Continuously conducting cluster analysis on the daily files of administrators can dynamically update the responsibility sets of leaf nodes and the association between administrators and responsibility sets. When the enterprise develops new business and generates new daily texts, the system can automatically identify and incorporate them into the cluster analysis, re-match administrators A and B, ensuring that the system always adapts to the dynamic changes of the enterprise and maintains the effectiveness of feedback processing and management process optimization.

[0025] In addition to the purposes, features and advantages described above, the present invention has other purposes, features and advantages. The present invention will be further described in detail below with reference to the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is the overall module schematic diagram of the present invention;

[0027] Figure 2 is the working principle flow chart of the intelligent management responsibility sorting module of the present invention;

[0028] Figure 3 is the schematic diagram of the intelligent feedback information positioning module of the present invention;

[0029] Figure 4 is the first schematic diagram of the hierarchical association feedback module of the present invention;

[0030] Figure 5 is the second schematic diagram of the hierarchical association feedback module of the present invention.

[0031] The meanings of each label in the figure are:

[0032] 100, Intelligent Management Responsibility Sorting Module; 200, Intelligent Feedback Information Location Module; 300, Hierarchical Association Feedback Module. Detailed Implementation Manner

[0033] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0034] Embodiment 1, refer to Figures 1 - 5 As shown, an enterprise management system based on artificial intelligence technology includes an intelligent management responsibility sorting module 100, an intelligent feedback information location module 200, and a hierarchical association feedback module 300. The detailed working principle is as follows:

[0035] Step 1: The intelligent management responsibility sorting module 100 perceives the daily texts processed by each administrator daily (the daily texts are files involving a large amount of information processing and decision-making in the daily operation of the enterprise and the work of the administrator). Cluster analysis is performed on multiple daily texts, and the daily texts with similarity are classified into one category, and the clustering results are represented by a tree diagram. Then, analyze the responsibility set of each leaf node in the tree diagram, determine the administrator corresponding to the leaf node, and establish the association relationship between each administrator and the responsibility set;

[0036] To implement the cluster analysis in the intelligent management responsibility sorting module 100: Perceive the responsibility field vocabulary in the enterprise organizational structure (the organizational structure is the relationship structure between each administrator within the enterprise, which clarifies the responsibilities, powers, and mutual relationships of different departments and positions within the enterprise), and construct a vocabulary dictionary with the responsibility field vocabulary; Traverse each daily text, compare each responsibility field vocabulary in the vocabulary dictionary one by one. If a responsibility field vocabulary in the vocabulary dictionary appears in the daily text, then add the number of occurrences of the responsibility field vocabulary in the daily text ; And according to the order of the responsibility field vocabulary in the vocabulary dictionary, take the number of occurrences of each responsibility field vocabulary in the daily text as an element in the daily text vector, and construct the daily text vector corresponding to each daily text. The specific expression is as follows:

[0037] Perceive the vocabulary dictionary as where is the total number of responsibility field vocabulary in the vocabulary dictionary, represents the th responsibility field vocabulary, perceive a daily text where is the total number of vocabulary in the duty field in daily text, represents the th word in the text;

[0038] Count the number of occurrences of a single duty field vocabulary: For any duty field vocabulary in the vocabulary dictionary , its number of occurrences in the daily text is calculated by the formula: where is an indicator function, when, ; when when, ; Traverse each duty field vocabulary in the daily text, and when it is judged to be the same as the duty field vocabulary, the count is accumulated;

[0039] Construct a text vector: Sense the number of occurrences of duty field vocabulary in the daily text , form a daily text vector , because, so each administrator's can be determined according to the duty field vocabulary in the daily text

[0040] The specific clustering analysis is a bottom-up clustering method. Starting from each daily text vector as a separate class, similar classes are continuously merged until a class containing all daily text vectors is formed; Starting from each daily text vector as a separate cluster, there are clusters, each cluster contains a daily text vector, The number of is equal to the daily text vector data; For each pair of clusters, calculate the similarity between each cluster to obtain a similarity matrix, and the similarity matrix records the similarity between each two cluster pairs. The corresponding expression is as follows:

[0041] Sense that the two daily text vectors are respectively and , where is the total number of duty field vocabulary in the vocabulary dictionary, and are respectively the number of occurrences of the th duty field vocabulary in the vocabulary dictionary in the daily text and the daily text , the daily text and the daily text similarity The calculation formula is:

[0042] ;

[0043] In the process of clustering analysis, the present invention also takes into account that, due to the obvious differences in the work content and responsibility scopes of different positions and administrators in an enterprise (for example, the daily texts of administrators in the sales department mainly revolve around customer development, order signing, etc., while the texts of administrators in the financial department focus on budget management, cost accounting, etc.), the intelligent management responsibility sorting module 100 also sets a similarity threshold to merge two clusters in the similarity matrix with similarity > similarity threshold, and merge the text vectors to form a more representative cluster. After merging, a new cluster is formed, and the new cluster contains all the text vectors in the original two clusters, so as to cluster the text vectors with similar work content together, more clearly understand the responsibility boundaries and work focuses of different administrators, and provide strong support for subsequent management decisions.

[0044] Repeat the above steps of calculating similarity and merging clusters. Each iteration will reduce one cluster until there is no cluster in the similarity matrix with similarity > similarity threshold; by setting the similarity threshold and merging similar clusters, similar daily text vectors can be grouped into one category, enabling administrators to be more targeted when processing information, so that subsequent administrators can focus on a category of responsibility areas represented by each cluster, rather than being overwhelmed by a large amount of messy feedback information, thereby improving the efficiency of information acquisition and processing, timely discovering and solving problems, and enhancing the overall operation efficiency of the enterprise.

[0045] The leaf nodes of the tree diagram in the intelligent management responsibility sorting module 100 are daily text vectors; each time two or more similar clusters are merged into a new cluster according to the similarity threshold, the new cluster is represented by a non-leaf node in the tree diagram. Each non-leaf node records an operation of merging clusters, and the non-leaf node contains information of all daily text vectors inherited from the child nodes; each non-leaf node represents an operation of merging clusters. The branches of the tree diagram represent the connection between different nodes, and the length of the branch represents the similarity between the two clusters during merging. The higher the similarity, the shorter the branch length, indicating that the two clusters are more similar in content; on the contrary, the lower the similarity, the longer the branch, indicating that the difference between the two clusters is relatively large. In this way, the tree diagram not only shows the clustering structure but also intuitively reflects the similarity degree between different clusters.

[0046] The intelligent management responsibility sorting module 100 retrieves the daily text corresponding to each leaf node in the tree diagram and each responsibility area vocabulary in each daily text, constructs a responsibility area vocabulary set from the responsibility area vocabularies in all leaf nodes, matches and calculates the occurrence frequencies of different responsibility area vocabularies in the responsibility area vocabulary set, removes duplicates from the responsibility area vocabulary set according to the occurrence frequencies of each responsibility area vocabulary. When the same vocabulary has different occurrence frequencies in different leaf nodes, the vocabulary with the maximum occurrence frequency is saved to the corresponding leaf node, and the responsibility area vocabulary set after deduplication is used as the responsibility set of the leaf node. The corresponding expression is as follows:

[0047] The set of words in the perceived responsibility area is , and the words in it occurrence frequency , where is a word in the responsibility area in the set of words in the responsibility area the number of occurrences; is the total number of words in the set of words in the responsibility area ;

[0048] For each leaf node , its set of responsibilities is selected from the set of words in the responsibility area the words with the highest occurrence frequency for the leaf node to form a set ; specifically, by traversing the leaf nodes and the set of words in the responsibility area, for each word , compare its occurrence frequency in different leaf nodes, and retain the word corresponding to the maximum frequency in the set of responsibilities of the corresponding leaf node. For the word , if the occurrence frequency in the leaf node is the maximum occurrence frequency among all leaf nodes for , then ; thus, after deduplication and determination of the set of responsibilities, the responsibilities between different leaf nodes have better comparability and analyzability, and it is possible to more intuitively see the differences and similarities in responsibilities of the daily texts represented by each leaf node, which is conducive to further mining the relationships between administrator responsibilities by the hierarchical association feedback module 300.

[0049] The working principle of the management responsibility intelligent sorting module 100 to establish the association relationship between each administrator and the set of responsibilities is as follows: Perceive the daily text corresponding to each leaf node and each administrator to which the daily text belongs, and retrieve all administrators whose daily text is only in one leaf node, defined as administrator A, and associate the set of responsibilities of the leaf node with administrator A; define the administrators in multiple leaf nodes as administrator B; associate the sets of responsibilities of multiple leaf nodes with administrator B, and its expression form is as follows: The perceived set of leaf nodes is , where represents the th leaf node; the set of daily texts corresponding to each leaf node is ;

[0050] The perceived set of administrators is , where represents the th administrator;

[0051] Let A be the set of Administrator A, B be the set of Administrator B, and the set of responsibilities of each leaf node is ;

[0052] For each administrator , calculate the number of leaf nodes where their daily texts are located . If , that is, there is a unique leaf node such that all the daily texts processed , then ; if , that is, there are multiple leaf nodes such that the texts in are distributed among the leaf nodes corresponding to the daily text set: , then ;

[0053] For administrator , let the corresponding leaf node be , then establish an association relationship , indicating that the set of responsibilities of administrator is ; for administrator , let the corresponding multiple leaf nodes be , then establish an association relationship , indicating that the set of responsibilities of administrator covers the sets of responsibilities of these leaf nodes. In an enterprise, the scope of responsibilities of different administrators varies. Some administrators focus on a single type of work, while some need to cover multiple fields. Thus, an administrator whose daily texts are only in one leaf node is defined as Administrator A, meaning that such an administrator's work is more concentrated and the responsibilities are clearly single. Associating the set of responsibilities of this leaf node with it can accurately reflect their actual work content. And an administrator who is in multiple leaf nodes is defined as Administrator B, indicating that such an administrator's work involves multiple different fields. Associating the sets of responsibilities of multiple leaf nodes with it also conforms to their actual responsibility distribution.

[0054] Step 2: The feedback information intelligent positioning module 200 senses the feedback text, and uses the method of extracting key responsibility area vocabulary from the daily text in the management responsibility intelligent sorting module 100 to extract the responsibility vocabulary in the feedback text. Compare the responsibility vocabulary in the feedback text with the set of responsibilities of each leaf node respectively to match the leaf node corresponding to the responsibility vocabulary in the feedback text:

[0055] If the duty vocabulary exactly matches the duty set of a leaf node, then retrieve the corresponding administrator A of the leaf node and output the feedback text to administrator A; if the duty vocabulary matches the duty sets of multiple leaf nodes respectively, then retrieve the corresponding administrator B of the leaf node and output the feedback text to administrator B. The corresponding expression is as follows:

[0056] Perceive the leaf node whose duty set is , where represents the th duty vocabulary in this duty set. For the duty vocabulary set of the feedback text and the duty set of the leaf node , if , then find the corresponding administrator A of the leaf node and output the feedback text to administrator A; if there are multiple leaf nodes such that , then find the corresponding administrator B of the leaf node and output the feedback text to administrator B.

[0057] Embodiment 2

[0058] Based on Embodiment 1, considering that the business scope of an enterprise is extremely complex, covering many fields such as production, sales, finance, human resources, etc., it is necessary to set up administrators at different levels to build a multi-level management system to achieve efficient operation. However, in the actual management process of an enterprise, due to factors such as dynamic business changes, departmental duty adjustments, and new business expansions, the organizational structure corresponding to the pre-set administrators at different levels in the enterprise management system often fails to match the actual organizational structure of the enterprise at present. Referring to Figure 4 and Figure 5 shown, the difference between this embodiment and Embodiment 1 is:

[0059] The hierarchical association feedback module 300 senses the administrators corresponding to each leaf node. If both administrator A and administrator B are included in the leaf nodes, it is determined that administrator B is the superior of administrator A. This is beneficial because when the same feedback text needs to be processed by administrator B, due to the heavy daily workload and numerous affairs of administrator B, although administrator B can judge based on the content of the feedback text that it needs to be transferred to the specific administrator A for processing, however, if a list of subordinate administrators A available for selection cannot be clearly presented to administrator B at this time, it will be difficult for administrator B to quickly and accurately assign the feedback text to the most suitable subordinate administrator A, thus reducing the feedback processing efficiency and affecting the enterprise's response speed to problems. Therefore, the hierarchical association feedback module 300 senses multiple leaf nodes with matching duty vocabulary, respectively retrieves administrator A of each leaf node, and when the feedback text is output to administrator B, administrator A and the feedback text are synchronously output to administrator B, enabling administrator B to quickly and accurately assign the feedback text without spending a lot of time and effort to find a suitable subordinate administrator A by himself, greatly shortening the transfer time of feedback processing and improving the overall efficiency of feedback processing.

[0060] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. Enterprise management system based on artificial intelligence technology, characterized by: It includes a management responsibility intelligent sorting module (100), a feedback information intelligent positioning module (200) and a hierarchical association feedback module (300), wherein: The management responsibility intelligent sorting module (100) performs cluster analysis on multiple daily texts, classifies daily texts with similarity into one category, and uses a tree diagram to represent the clustering results, analyzes the responsibility set of each leaf node in the tree diagram, determines the administrator corresponding to the leaf node, and establishes an association relationship between each administrator and the responsibility set; the feedback information intelligent positioning module (200) perceives the responsibility vocabulary in the feedback text, matches the responsibility set of the leaf node corresponding to the responsibility vocabulary in the feedback text, and after matching, calls out the administrator corresponding to the leaf node to push the feedback text; the hierarchical association feedback module (300) determines the superiors and subordinates of the administrator based on the tree diagram, and when pushing the feedback text to the superior administrator, calls out the subordinate administrator corresponding to the feedback text for synchronous push.

2. The enterprise management system based on artificial intelligence technology according to claim 1, characterized in that: The working steps of cluster analysis in the management responsibility intelligent sorting module (100) are as follows: Step 1: Perceive the responsibility domain vocabulary in the corporate organizational structure and construct the responsibility domain vocabulary into a vocabulary dictionary; traverse each daily text and compare the responsibility domain vocabulary in the vocabulary dictionary one by one. If the responsibility domain vocabulary in the vocabulary dictionary appears in the daily text, the number of occurrences of the responsibility domain vocabulary in the daily text is increased. ; According to the order of the vocabularies in the vocabulary dictionary, the number of times each vocabularies in the vocabularies appear in the daily text is used as an element in the daily text vector to construct the daily text vector corresponding to each daily text; Step 2: Cluster analysis is a bottom-up clustering method that starts with each daily text vector as a separate class and continuously merges similar classes until a class containing all daily text vectors is formed; each daily text vector is taken as a separate cluster, with a total of clusters, each containing a daily text vector, The number is equal to the daily text vector data; for each pair of clusters, the similarity between each cluster is matched and calculated to obtain a similarity matrix, which records the similarity between each pair of clusters; Set a similarity threshold, merge the two clusters whose similarity in the similarity matrix is ​​greater than the similarity threshold, merge the text vectors to form a more representative cluster, and form a new cluster after merging. The new cluster contains all the text vectors in the original two clusters. Repeat the above steps of calculating similarity and merging clusters, reducing one cluster in each iteration until no cluster with similarity greater than the similarity threshold appears in the similarity matrix.

3. The enterprise management system based on artificial intelligence technology according to claim 2 is characterized in that: The management responsibility intelligent combing module (100) matches and calculates the similarity between each cluster: it perceives that two daily text vectors are respectively and ,in is the total number of domain-specific words in the vocabulary dictionary, and The first The vocabulary of the responsibility area in daily text and daily text The number of occurrences in daily text and daily text Similarity The calculation formula is: 。 4. The enterprise management system based on artificial intelligence technology according to claim 1, characterized in that: The leaf nodes of the dendrogram in the management responsibility intelligent combing module (100) are daily text vectors; each time two or more similar clusters are merged into a new cluster according to a similarity threshold, the new cluster is represented by a non-leaf node in the dendrogram, and each non-leaf node records a cluster merging operation; each non-leaf node represents a cluster merging operation, and the branches of the dendrogram represent connections between different nodes, and the length of the branches represents the similarity between the two clusters when merged.

5. The enterprise management system based on artificial intelligence technology according to claim 4 is characterized in that: The management responsibility intelligent combing module (100) calls out the daily text corresponding to each leaf node in the tree diagram, and each responsibility field word in each daily text, constructs the responsibility field words in all leaf nodes into a responsibility field word set, matches and calculates the frequency of occurrence of different responsibility field words in the responsibility field word set, removes duplicates from the responsibility field word set according to the frequency of occurrence of each responsibility field word, and when the frequency of occurrence of the same word in different leaf nodes is different, the word with the maximum frequency of occurrence is saved in the corresponding leaf node, and the responsibility field word set after deduplication is used as the responsibility set of the leaf node.

6. The enterprise management system based on artificial intelligence technology according to claim 5, characterized in that: The working principle of the management responsibility intelligent sorting module (100) for establishing the association relationship between each administrator and the responsibility set is as follows: perceive the daily text corresponding to each leaf node and each administrator to whom the daily text belongs, call out all administrators whose daily texts are only in one leaf node, define them as administrator A, and associate the responsibility set of the leaf node with administrator A; define the administrator in multiple leaf nodes as administrator B; and associate the responsibility sets of multiple leaf nodes with administrator B.

7. The enterprise management system based on artificial intelligence technology according to claim 1, characterized in that: The feedback information intelligent positioning module (200) adopts the method of extracting key responsibility domain words in daily text in the management responsibility intelligent combing module (100), extracts responsibility words in the feedback text, compares the responsibility words in the feedback text with the responsibility set of each leaf node, and matches the leaf nodes corresponding to the responsibility words in the feedback text: If the responsibility vocabulary completely matches the responsibility set of a leaf node, the administrator A corresponding to the leaf node is called out and the feedback text is output to administrator A; if the responsibility vocabulary matches the responsibility sets of multiple leaf nodes respectively, the administrator B corresponding to the leaf node is called out and the feedback text is output to administrator B.

8. The enterprise management system based on artificial intelligence technology according to claim 7 is characterized in that: The feedback information intelligent positioning module (200) senses the leaf node The set of responsibilities is ,in Represents the first Responsibility vocabulary, a collection of responsibility vocabulary for feedback text and leaf nodes Responsibilities ,like , then find the leaf node The corresponding administrator A outputs the feedback text to administrator A; if there are multiple leaf nodes , so that , then find the administrator B corresponding to the leaf node and output the feedback text to administrator B.

9. The enterprise management system based on artificial intelligence technology according to claim 6, characterized in that: The hierarchical association feedback module (300) senses the administrator corresponding to each leaf node, and if the leaf nodes both contain administrator A and administrator B, it is determined that administrator B is the superior of administrator A.

10. The enterprise management system based on artificial intelligence technology according to claim 9, characterized in that: The hierarchical association feedback module (300) senses multiple leaf nodes matched by responsibility words, and respectively calls out the administrator A of each leaf node. When the feedback text is output to the administrator B, the administrator A and the feedback text are output to the administrator B synchronously.

Images