CN113742496B - A power knowledge learning system and method based on heterogeneous resource fusion - Google Patents

Abstract

The present application discloses an electric power knowledge learning system and method based on heterogeneous resource fusion, the system includes a resource layer, a business layer and an application layer; the resource layer is used to obtain electric power resource information; the business layer is used to construct an electric power knowledge base and an electric power knowledge graph based on the electric power resource information, and is used to construct portraits of employees in different positions based on employee information; the application layer is used to provide personalized services for each employee, and the personalized services include electric power knowledge recommendation and electric power knowledge test paper evaluation. The present application can provide targeted training for each employee, recommend corresponding electric power knowledge according to the job requirements and preference characteristics of each employee, and can also provide personalized electric power test questions for each employee and conduct evaluation; the present application can also sort out the messy electric power knowledge, and can clearly see the relationship between electric power knowledge, thereby improving the learning efficiency of power grid staff.

Description

A power knowledge learning system and method based on heterogeneous resource fusion

Technical Field

The present application relates to the technical field of power system construction, and mainly relates to a power knowledge learning system and method based on heterogeneous resource fusion.

Background technique

Experiences from countries around the world show that the development speed of electricity production should be higher than that of other sectors in order to promote the coordinated development of the national economy. Therefore, the power industry is also known as the "pioneer" of the national economy. Due to the high requirements for the development of the power industry, it is also crucial to improve the quality and ability of each power grid worker. Therefore, it is necessary to provide power knowledge training to each power grid worker.

Nowadays, most of the training for power grid workers is carried out through training lectures held by enterprises, one-to-many training by professional teachers, or self-study by grid workers by searching for information on the Internet. However, the resources of knowledge in the power industry are extensive and huge in number. When searching, each worker cannot filter out the resources he needs from a large amount of data. In addition, the demands of grid workers in each position are different, and the objects they need to focus on are different. One-to-many training cannot achieve targeted improvement based on the characteristics of each grid worker. However, if a professional teacher is assigned to each grid worker, this will consume a lot of resources and the cost will be too high.

Summary of the invention

Since the existing technology cannot provide targeted electricity knowledge training to each power grid worker, the present application discloses an electricity knowledge learning system and method based on heterogeneous resource fusion.

The first aspect of the present application discloses an electric power knowledge learning system based on heterogeneous resource fusion, including a resource layer, a business layer and an application layer; the resource layer is used to obtain electric power resource information; the business layer is used to construct an electric power knowledge base and an electric power knowledge graph according to the electric power resource information, and is used to construct portraits of employees in different positions according to employee information; the application layer is used to provide personalized services for each employee, and the personalized services include electric power knowledge recommendations and electric power knowledge test paper evaluations.

The second aspect of the present application discloses a method for learning electric power knowledge based on heterogeneous resource fusion, the method being applied to the electric power knowledge learning system based on heterogeneous resource fusion, comprising:

Obtain power resource information;

Build a knowledge base on electricity;

Build a power knowledge graph;

Build portraits of employees in different positions;

Provide personalized services for each employee; the personalized services include power knowledge recommendations and power knowledge test paper evaluation.

Optionally, the acquiring of power resource information includes:

Acquire original information of electric power resources, wherein the original information of electric power resources includes: original electric power audio information, original electric power video information and original electric power document information;

Acquire power resource information, the power resource information including: power audio information, power video information and power document information; the power audio information is to unify the format of the original power audio information; the power video information is to unify the format of the original power video information; the power document information is to unify the format of the original power document information, and extract text information from the power audio information and the power video information, and unify the format of the text information.

Optionally, the building of the electric power knowledge base includes:

Extracting keyword information from the electric power resource information;

Acquire classified power resource information; the classified power resource information includes: classifying the keyword information, and summarizing the power resource information corresponding to each type of keyword information;

Each type of power resource information is processed to construct a power knowledge base; the processing includes content analysis, disambiguation of identical words, removal of duplicate content, and fusion of similar content.

Optionally, the constructing of the electric power knowledge graph includes:

Based on the power knowledge base, extract power professional vocabulary and build a power professional vocabulary list;

Based on the electric power professional vocabulary and each type of electric power resource information, obtaining the connection between the electric power professional vocabulary;

Each electric power professional vocabulary is taken as a label, and the connection between the electric power professional vocabulary is taken as an edge to construct an electric power knowledge graph.

Optionally, the constructing of employee portraits in different positions includes:

Get information about employees in different positions;

Extract features from employee information of each position and build a profile of employees in each position;

Obtain portraits of employees in different positions.

Optional, provide personalized services for each employee, including:

Analyze the needs of employees in different positions based on the portraits of employees in different positions;

Obtaining the preference characteristics of each employee based on the historical browsing information of each employee in the power knowledge graph;

Based on the information each employee is currently looking for, each employee's preference characteristics, and each employee's corresponding job requirements, electricity knowledge recommendations are made for each employee.

Optionally, after recommending power knowledge to each employee based on the information currently being searched by each employee, the preference characteristics of each employee, and the job requirements corresponding to each employee, the method further includes:

Obtaining feedback information from each employee on the power knowledge recommendation;

Based on the feedback information, the next recommended content of electricity knowledge for each employee is optimized.

Optional, personalized services are provided for each employee, including:

The electric power knowledge base includes electric power knowledge test questions;

According to the preference characteristics of each employee and the corresponding job requirements of each employee, electric power knowledge test questions are selected from the electric power knowledge base for each employee, and the selected electric power knowledge test questions are combined into an electric power knowledge test paper;

Score the electricity knowledge test paper completed by each employee and obtain each employee's score.

Optional, personalized services are provided for each employee, including:

Obtaining feedback information from each employee on the electricity knowledge test paper;

Based on the feedback information and the score of each employee, the content of the next electricity knowledge test paper for each employee is optimized.

The present application discloses an electric power knowledge learning system and method based on heterogeneous resource fusion, the system comprising a resource layer, a business layer and an application layer; the resource layer is used to obtain electric power resource information; the business layer is used to construct an electric power knowledge base and an electric power knowledge graph according to the electric power resource information, and is used to construct portraits of employees in different positions according to employee information; the application layer is used to provide personalized services for each employee, and the personalized services include electric power knowledge recommendations and electric power knowledge test paper evaluations.

This application can provide targeted training for each employee, recommend corresponding electricity knowledge based on each employee's job requirements and preference characteristics, and provide personalized electricity test questions and conduct assessments for each employee; this application can also construct the huge and complex knowledge in the electricity field into an electricity knowledge base and electricity knowledge graph, sort out the messy electricity knowledge, clearly see the relationship between electricity knowledge, and improve the learning efficiency of power grid staff.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solution of the present application, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

FIG1 is a schematic diagram of the structure of a power knowledge learning system based on heterogeneous resource fusion disclosed in an embodiment of the present application;

FIG2 is a flow chart of a method for learning electric power knowledge based on heterogeneous resource fusion disclosed in an embodiment of the present application.

Detailed ways

Since the existing technology cannot provide targeted electricity knowledge training to each power grid worker, the present application discloses an electricity knowledge learning system and method based on heterogeneous resource fusion.

The first embodiment of the present application discloses an electric power knowledge learning system based on heterogeneous resource fusion, referring to the structural diagram shown in FIG1 , which includes a resource layer, a business layer and an application layer; the resource layer is used to obtain electric power resource information; the business layer is used to construct an electric power knowledge base and an electric power knowledge graph according to the electric power resource information, and to construct employee portraits of different positions according to employee information; the application layer is used to provide personalized services for each employee, and the personalized services include electric power knowledge recommendation and electric power knowledge test paper evaluation. Specifically, a basic service platform is built using technologies such as spark or hadoop, and the acquired electric power resource information is stored in the hardware basic environment and equipment.

The second embodiment of the present application discloses a method for learning electric power knowledge based on heterogeneous resource fusion, and the method is applied to the electric power knowledge learning system based on heterogeneous resource fusion, referring to the flow chart shown in FIG2 , including:

Obtain original information on electric power resources, which includes original electric power audio information, original electric power video information and original electric power document information; wherein, there are many ways to obtain the original electric power document information, including internal learning resources and external learning resources, wherein the online college, document center and digital library constitute internal learning resources; the palmtop college and information website constitute external learning resources; divide the content and category of resources, collect and organize resources from different carriers, and for intranet resources, connect learning resources from the human resources system, online college, palmtop college, digital library and document center system, combine the storage methods in different systems, build an intranet resource integration platform, develop an intranet connection interface, use crawler technology to capture external resources on the external information website, copy them to the intranet system, and then distinguish categories to establish a resource storage repository.

Obtain power resource information, the power resource information includes: power audio information, power video information and power document information; the power audio information is to unify the format of the original power audio information; the power video information is to unify the format of the original power video information; the power document information is to unify the format of the original power document information, and extract the text information in the power audio information and the power video information, and unify the format of the text information. The formats of the acquired power knowledge resources are varied and need to be parsed and unified for easy use. Since audio files and video files are one of the main forms of power knowledge, the text information in the audio and video can be extracted, and a large amount of obscure professional content and professional terms often appear in the audio and video, resulting in low recognition accuracy. A corresponding speech training model in the power field can be constructed to improve the accuracy of speech recognition.

The original power document information obtained has great differences in structure. The mainstream formats are txt, word, ppt, pdf, epub, etc. The storage format and content structure of different types of resources are completely different. It is impossible to read and process all formats of content in a universal way. Therefore, it is necessary to design and provide an adaptive framework to solve the problem of reading (recognizing) diverse and heterogeneous files. The adapter mode is an extensible and highly customized solution. This application includes a variety of parsers. For different document structures, dedicated parsers are provided, such as epub parser, pdf parser, mp4 video parsing processor, etc., and they are distributed to the required processing units through adapters in the processing project, so that the entire system has both high integration and unity and high adaptability and extensibility.

The original document information is processed logically in a structured way to extract the content from heterogeneous and noisy data. Usually, the text content consists of multiple sentences, and there is a certain logical structural relationship between sentences, such as the chapter structure, paragraph structure, and previous and next sentences of a book. Some storage format files have structural information themselves, such as epub format files, which contain chapter information of the content in the toc file; in pdf files, the outline of the document is also stored in the Outline Tree part. Due to the technical level and ability of the document producer, the structure itself is often neither accurate nor comprehensive, and more document formats, such as txt and doc files, do not have the concept of structure and can often only be presented through calibrated words or formats. Therefore, through the logical processing of the content, the content is structured, which can quickly locate the content (sentences) on the one hand, and provide a basis for subsequent semantic processing on the other hand.

Extracting keyword information from the power resource information;

Acquire classified power resource information; the classified power resource information includes: classifying the keyword information, and summarizing the power resource information corresponding to each type of keyword information;

Each type of power resource information is processed to construct a power knowledge base; the processing includes content analysis, disambiguation of identical words, removal of duplicate content, and fusion of similar content. Specifically, implicit knowledge inference technology is used to establish associations between knowledge points in the multi-category knowledge system in the knowledge center, and a network system is established for the scattered content in the knowledge center. After the network system is established, knowledge discovery technology is used to timely update and expand the knowledge system, and then knowledge disambiguation technology is used to remove redundant knowledge items. Finally, knowledge fusion technology is used to form a knowledge center with a clear structure and distinct levels.

Based on the power knowledge base, extract power professional vocabulary and build a power professional vocabulary list;

Based on the electric power professional vocabulary and each type of electric power resource information, obtaining the connection between the electric power professional vocabulary;

Each electric power professional vocabulary is taken as a label, and the connection between the electric power professional vocabulary is taken as an edge to construct an electric power knowledge graph.

After the knowledge resources are parsed and identified, the unstructured, polymorphic documents or contents are converted into objects with a certain structure. Each object is assigned a unique identifier, and the relationship between objects is established, thus laying the foundation for subsequent processing calculations and content association fusion. For example, within the same training resource, the order of training-related content or topics, chapter relationships, etc. can be established; different training resources can be centered around resource-extracted topics, knowledge points, or external related attributes such as departments, authors, events, etc., to establish author relationship chains for journal articles and documents, relationships between question banks and training materials, case events, etc., so that scattered training content and materials can be integrated to form a knowledge system.

Different types of knowledge correspond to points of different colors, or points with different labels. Similarly, different types of relationships correspond to edges of different colors, or edges with different labels. If the relationship is bidirectional, the corresponding edge is a bidirectional edge, otherwise, it corresponds to a unidirectional edge. The edges corresponding to relationships of different strengths have different weight values.

Get information about employees in different positions;

Extract features from employee information of each position and build a profile of employees in each position;

Obtain portraits of employees in different positions.

Build portraits of employees in different positions;

Analyze the needs of employees in different positions based on the portraits of employees in different positions;

Obtaining the preference characteristics of each employee based on the historical browsing information of each employee in the power knowledge graph;

According to the information each employee is currently looking for, the preferences of each employee, and the corresponding job requirements of each employee, we recommend electricity knowledge to each employee. We recommend the most similar electricity knowledge to employees for learning.

Obtaining feedback information from each employee on the power knowledge recommendation;

Based on the feedback information, the next recommended content of electricity knowledge for each employee is optimized.

The electric power knowledge base includes electric power knowledge test questions;

According to the preference characteristics of each employee and the corresponding job requirements of each employee, electric power knowledge test questions are selected from the electric power knowledge base for each employee, and the selected electric power knowledge test questions are formed into an electric power knowledge test paper;

Score the electricity knowledge test paper completed by each employee and obtain each employee's score.

Obtaining feedback information from each employee on the electricity knowledge test paper;

Based on the feedback information and the score of each employee, the content of the next electricity knowledge test paper for each employee is optimized.

For the power knowledge test paper obtained in this application, the web crawler processing technology is used, and the specific operations for this part are as follows:

1) First, select a part of the data seed URL;

2) Put these URLs into the URL queue to be crawled;

3) Take the URL to be crawled from the queue of URLs to be crawled, and use the online DNS resolver to complete the DNS resolution process, get the IP address of the host, and download the web page corresponding to the URL, set it in a specific file area, or directly store it in the corresponding network cloud, and then store it in the downloaded web page library, and then put these URLs into the queue of crawled URLs;

4) Analyze the URLs in the crawled URL queue, analyze other URLs in it simultaneously, and put these URLs into the queue of URLs to be crawled, thus entering the next cycle.

The present application discloses an electric power knowledge learning system and method based on heterogeneous resource fusion, the system comprising a resource layer, a business layer and an application layer; the resource layer is used to obtain electric power resource information; the business layer is used to construct an electric power knowledge base and an electric power knowledge graph according to the electric power resource information, and is used to construct portraits of employees in different positions according to employee information; the application layer is used to provide personalized services for each employee, and the personalized services include electric power knowledge recommendations and electric power knowledge test paper evaluations.

This application can provide targeted training for each employee, recommend corresponding electricity knowledge based on each employee's job requirements and preference characteristics, and provide personalized electricity test questions and conduct assessments for each employee; this application can also construct the huge and complex knowledge in the power field into an electricity knowledge base and an electricity knowledge graph, sort out the messy electricity knowledge, clearly see the relationship between electricity knowledge, and improve the learning efficiency of power grid staff.

The present application is described in detail above in conjunction with specific implementation methods and exemplary examples, but these descriptions cannot be understood as limitations on the present application. Those skilled in the art understand that, without departing from the spirit and scope of the present application, a variety of equivalent replacements, modifications or improvements can be made to the technical solutions and implementation methods of the present application, all of which fall within the scope of the present application. The scope of protection of the present application shall be subject to the attached claims.

Claims (7)

1. The electric power knowledge learning method based on heterogeneous resource fusion is characterized in that the method is applied to an electric power knowledge learning system based on heterogeneous resource fusion, a spark or hadoop technology is used for building a basic service platform, external resources are copied to an intranet system, and a resource storage library is built in a classified mode;

The system comprises a resource layer, a service layer and an application layer; the resource layer is used for acquiring power resource information; the business layer is used for constructing an electric power knowledge base and an electric power knowledge map according to the electric power resource information and constructing staff portraits of different posts according to staff information; the application layer is used for providing personalized services for each employee, wherein the personalized services comprise electric power knowledge recommendation and electric power knowledge test paper evaluation;

The method comprises the following steps:

Acquiring power resource information;

Constructing a power knowledge base;

constructing a power knowledge graph;

Constructing staff portraits of different posts;

Providing personalized services for each employee; the personalized service comprises power knowledge recommendation and power knowledge test paper evaluation;

The constructing the power knowledge base comprises:

Extracting keyword information aiming at the power resource information;

Acquiring classified power resource information; the classified power resource information includes: classifying the keyword information and summarizing the power resource information corresponding to each type of keyword information;

establishing a correlation relationship between knowledge points in a multi-category knowledge system through an implicit knowledge inference technology, establishing a network system, and removing redundant knowledge items by using a knowledge disambiguation technology to form a knowledge center with clear structure and clear hierarchy;

processing each type of power resource information to construct a power knowledge base; the processing includes content analysis, identical word disambiguation, duplicate content removal, and similar content fusion;

the construction of the power knowledge graph comprises the following steps:

aiming at the electric power knowledge base, extracting electric power professional vocabulary and constructing an electric power professional vocabulary list;

Acquiring a relation between the power professional vocabularies based on the power professional vocabulary and the power resource information of each type;

taking each electric power professional vocabulary as a label, and taking the relation between the electric power professional vocabularies as an edge to construct an electric power knowledge graph;

And analyzing and identifying the knowledge resources, converting the document or the content into objects with a certain structure, giving unique identifiers, and establishing the connection between the objects.

2. The method for learning power knowledge based on heterogeneous resource fusion according to claim 1, wherein the step of obtaining power resource information comprises:

Acquiring power resource original information, wherein the power resource original information comprises: original power audio information, original power video information, and original power document information;

acquiring power resource information, the power resource information comprising: power audio information, power video information, and power document information; the power audio information is obtained by unifying the formats of the original power audio information; the power video information is obtained by unifying the formats of the original power video information; the power document information is obtained by unifying the formats of the original power document information, extracting text information in the power audio information and the power video information, and unifying the formats of the text information.

3. The power knowledge learning method based on heterogeneous resource fusion of claim 1, wherein the constructing staff portraits of different posts comprises:

Staff information of different posts is obtained;

Extracting the characteristics of staff information of each post, and constructing staff portraits of each post;

And obtaining staff images of different posts.

4. The power knowledge learning method based on heterogeneous resource fusion according to claim 1, wherein providing personalized services for each employee comprises:

analyzing the demands of staff at different positions according to the pictures of staff at different positions;

acquiring preference characteristics of each employee according to historical browsing information of each employee in the power knowledge graph;

And recommending the electric power knowledge for each employee according to the information to be searched currently by each employee, the preference characteristics of each employee and the corresponding post requirements of each employee.

5. The method for learning electric power knowledge based on heterogeneous resource fusion according to claim 4, wherein after recommending electric power knowledge for each employee according to the information currently searched by each employee, the preference characteristics of each employee, and the corresponding post requirements of each employee, the method further comprises:

Acquiring feedback information of each employee on the power knowledge recommendation;

And optimizing the next power knowledge recommendation content of each employee according to the feedback information.

6. The power knowledge learning method based on heterogeneous resource fusion of claim 4, wherein a personalized service is provided for each employee, further comprising:

the electric power knowledge base comprises electric power knowledge questions;

Aiming at the preference characteristics of each employee and the post requirements corresponding to each employee, selecting power knowledge test questions from the power knowledge base for each employee, and forming the selected power knowledge test questions into a power knowledge test paper;

and scoring the power knowledge paper completed by each employee to obtain the score of each employee.

7. The power knowledge learning method based on heterogeneous resource fusion of claim 6, wherein a personalized service is provided for each employee, further comprising:

Acquiring feedback information of each staff on the electric power knowledge test paper;

and optimizing the content of the next power knowledge test paper of each employee according to the feedback information and the score of each employee.