九年级上册英语翻译
I.J. Education and Management Engineering, 2020, 2, 1-10
Published Online April 2020 in MECS (s-press)
DOI: 10.5815/ijeme.2020.02.01
整理英文Available online s-press/ijem
Development of Knowledge Graph for University Cours
Management
Ismail Aliyu1*, A. F. D. Kana2 and Salisu Aliyu3
1，2，3Department of Computer Science, Ahmadu Bello University Zaria, Nigeriawindowed mode
Received: 10 January 2020; Accepted: 14 February 2020; Published: 08 April 2020
Abstract
The task of Allocating cours to lecturers in many tertiary institutions is done manually by typing using word processor application. Motivated by the widespread application of knowledge graphs in different domains, we prent automated approach bad on knowledge graph to address the problem of manual cour allocation to, a task usually carried out at the beginning of every mester or academic year by departments in tertiary institutions. The development of knowledge graph in a way that enables easy manipulation and automatic generation of cour allocation schedule is the core contribution of this paper. Rather than storing the data in relational databa tables, the system stores data in a knowledge graph which is in RDF/XML format and refer to it to support intelligent knowledge rvices. In addition to automatic generation of cour allocation schedule, another important feature of the system propod in this paper is its ability to enable easy implementation of tasks similar to Question Answering that are very important to education administrators, which the existing manual approach does not provide. Testing of the propod system reveals its ability to perform effectively. Our approach of using Knowledge graph offers advantages such as flexibility and curity.
Index Terms: Cour allocation, Knowledge graph, Resource Description Framework (RDF)
© 2020 Published by MECS Publisher. Selection and/or peer review under responsibility of the Rearch Association of Mode rn Education and Computer Science
* Corresponding author.
E-mail address:
1.Introduction and Literature Review
Allocation of cours to lecturers is an important task that is carried out at the beginning of every academic year or mester. However, in most Nigerian tertiary institutions it is done manually by simply typing on Microsoft word application. A simple question of who taught CS121 in cond mester of 2010 academic ssion?, for example, will amount to hours of arching for printout of cour allocation schedule of that mester/year or veral minutes of arching the computer for MS word file containing the information. In most cas, the word files stored on computer’s hard drive get corrupted or lost. Another problem is lack of automated approach for centralize tracking of what is going on at various departments with regard to allocation and teaching of cours. Considering the facts that number of cours taught by a lecturer is an important parameter that determines the lecturer’s Earned Academic Allowance (EAA), such centralize tracking systems can a
ssist school management or directorate of academic planning of the university for verification of academic staff claim for payment of EAA. Unfortunately, such systems do not exist in many Nigerian universities.
In recent past, Knowledge Graph (KG) has been introduced as a model for knowledge reprentation in a structured way. Knowledge graph reprents entities such as people, places, organizations etc and their relationships. Knowledge graphs have been incorporated in many applications in different domains to support various tasks. Previous rearch that applied knowledge graph in education domain such as [1] ud it to support learning and teaching of mathematics. [2] constructed knowledge graph to support scientific resource retrieval for students. To the best of our knowledge no rearch work was found that leverages knowledge graph to support routine activities related to management and teaching of cours in tertiary institutions. Our aim in this paper is to address the problem manual cour allocation using knowledge graph as a model of data storage. Specifically, we construct KG of cours and u the KG to support management of cours, automatic generation of cour allocation schedule, and implementation of Question Answering tasks that are important to education administrator. The contributions of this paper are as follows:
i．We apply knowledge graph to education domain to support management of cours and automati
c generation of cour allocation information in tertiary institutions.
ii．Propo an approach that enable education administrator to easily have access to basic information on teaching of cours at various departments. Answers to simple questions (that are often difficult to answer) can be obtained easily. – Examples; How many cours did lecturerY taught in a particular ssion or mester? Which cour is prerequisite to courX? What are the recommended reading materials for cour X?
The rest of this paper is organized as follows. Brief explanation on knowledge graph and some of its applications is the content of ction 2. Review of the literature is prented in ction 3. The method of building and querying the cours knowledge graph is described in ction 4. We describe the ca study in which the system was evaluated in ction 5. Section 6 concludes the paper.
2.KNOWLEDGE GRAPH
铍青铜英文Basically, Knowledge Graph is model for knowledge reprentation. It describes real world entities, objects, concepts etc and their relationships in a graph [3]. It is a structured knowledge ba that reprents knowledge as triple of the form (h,r,t). Where h is the head entity, t is the tail entity, and r i
s the relation between h and t. Example of knowledge graph is shown in fig.1. which has cours and lecturers as entities. The triple (Prof S.B, taught_2_2018, CS311) is interpreted as Prof. S.B taught CS311 in the cond mester of 2018 ssion. Similarly, the interpretation of the triple (CS121, prerequisite, CS211)is CS121 is the prerequisite of CS211.
Fig.1. Cours knowledge graph
Knowledge graph can be en as a network in which nodes are the entities and edges that link entiti
北京王府学校
es are the relations. Formally, Let E be the t of entities and R be the t of relations. Knowledge Graph, G is defined as collection of triple facts (e h, r, e t). That is, G = {(e h, r, e t) | e h,e t ∈ E , r∈R}.
Knowledge graph was first introduced by Google as a mantic enhancement of their arch functions, so that arch is beyond “keyword” matching [4]. Since its introduction in 2012, knowledge graph has promoted the development of “mantic” network of entities, which the giant tech companies (Google, Facebook, etc) and many enterpris u to support development of intelligent applications in their domains [5]. Knowledge graphs powered many Artificial Intelligence applications such as arch engines, Question Answering systems, Recommender system etc. They have also been applied to variety of domains like Education [1], Biomedical [6], Electric Power System [7], Automobile industry [8], Traditional Chine Medicine (TCM) to support intelligent knowledge rvices (such as efficient retrieval, recommendation etc) [9] and many more. Examples of publically available large scale knowledge graphs are DBpedia, YAGO, NELL, Freeba, and Word net.
Knowledge graphs can be curated manually or automatically. They can be constructed by automatically extracting instances of entities, and relations from structured, mi-structured or unstructured resources like text corpus or web. State of the art Natural Language Processing (NLP) t
echniques for gmentation, tagging, Named Entity Recognition (NER) are employed in the process of entity extraction. Another way to construct knowledge graph is to u data of specific domain aiming at solving some specific questions. Mostly the data are stored in relational databas.
3.RELATED WORKwe are never ever
Our aim in this paper is to construct knowledge graph for u in education domain to support management of cours in tertiary institutions. [1] propo system called KnowEdu that automatically construct educational knowledge graph to support teaching and learning of mathematics. KnowEdu us neural quence labelling algorithm on pedagogical data (eg, curriculum) to extract concepts of subjects or cours and employs probabilistic association rule mining to identify relations. The nodes of their KG are instructional concepts of subjects. An instructional concept is a basic concept that learner is expected to fully understands eg “linear equation” in Mathematics, “photosynthesis” in Biology. [10] construct a knowledge graph utilizing education
cherokee4Development of Knowledge Graph for University Cours Management
data mined from web, and develop visualization analysis platform called EduVis which supports ur
s to do associated analysis to reveal patterns and help education administrators take decisions bad on the data in the graph. [2] designed scientific publication knowledge graph to support scientific resource retrieval and other rvices for students. The scientific publications knowledge graph eks to integrate and link scientific publications with entities such as journal, rearchers, their affiliations etc, in order to enhance retrieval efficiency and reduce the difficulty of exploring scientific publications, thereby improving students’ and rearchers’ learning ability. The entities, some of which are metadata (such as, title, author, journal etc) were extracted from 3 scientific databas namely, Web of science, Engineering Village, and EBSCO. In educational domain, very few studies focus on construction of domain-specific knowledge graphs. However, some recent works investigated relation extraction between certain entities in educational domain. [11] create Prerequisite Structure Graph (PSG) using an unsupervid approach that utilizes text content and student activity log from heterogeneous sources. The nodes reprent the universal concepts in an educational domain and the edges specify the pairwi ordering of concepts in effective teaching by instructors or for effective learning by students. [12] recover prerequisite relations from university cour dependencies. [13] and [14] came up with a method called Concept Graph Learning (CGL) that automatically map online cours from different providers (universities or Massive Online Open Cours (MOOCs)) onto space of concepts, and predict latent prerequisite dependencies among both concepts and cours.
韩语发音器Like our work, the previous work reviewed above constructs knowledge graph and us it to support some tasks in education domain. In contrast, our work constructs and u knowledge graph to simplify the task of cour allocation and to support easy implementation Question Answering tasks that are important to education administrators. While some works [1] and [2] extract instance of entities from unstructured text, we u different approach instead. The instances of entities come from structured source such as relational databa table, an approach ud by [9]. Furthermore, we do not ud NLP techniques for Name Entity Recognition (NER) and relation extraction.
4.METHOD
This ction describes our method for constructing and querying the cours knowledge graph.
4.1 Data Schema
Knowledge graph is compod of entities of different types. The description of the types or class of entities and constraints on their u is usually given in a schema or ontology [3]. Figure 2 shows the schema of the cours knowledge graph. The entities are of two types; person and thing. Therefore the cours knowledge graph propod in this work do not include entities such as places (cities & countries), organizations etc that are usually included in conventional knowledge graphs.
The entities and relations ud in constructing cours knowledge graph are shown in Tables 1 and 2 respectively.
T ABLE 1.L IST OF ENTITIES OF COURSES KNOWLEDGE GRAPH
Entity Type Description
Lecturer (L) person Reprents a lecturer that teaches a cour
Cour (C) thing Reprent a cour that lecturer teaches
courBook (CB) thing Reprent the text book recommended for a particular cour
Bookauthor (BA) person The author’s name of the recommended text book
圣诞节英语对话Development of Knowledge Graph for University Cours Management 5
Fig.2. Data schema of the cours management model
T ABLE 2.L IST OF RELATIONS OF COURSES KNOWLEDGE GRAPH
Relation Description
Prerequisite relation between cour and another cour entities
taught_in_mester_year relation between cour entity and lecturer entity
writtenBy relation between author entity and courbook entity
mirror
referenceText relation between cour entity and courbook entity
4.2 Building the Cours Knowledge Graph
Fig.3. illustrates the architecture of the system that builds cours knowledge graph. The two main modules are Entity extraction, and Relation extraction. In this work, instances of entities are extracted from databa since list of cours and lecturers of a department are usually stored in databa. Relationships between cours and lecturers are established in an adhoc manner during
allocation of cours at the beginning of every mester or academic year. Becau of this, we create a Graphical Ur Interface (GUI) for ur to specify such relationships. When ur lects lecturer’s name, cour code, mester and year, such action is interpreted as relation of the form taught_in_mester_year. Conquently, a triple of the form (Cour, taught_in_mester_year, Lectuerer) is formed – eg (CS323, taught_in_1_2018, Dr S.A), interpreted as CS323 was taught in first mester of 2018 ssio by Dr S.A. Through the GUI, ur can equally specify recommended text book for a cour, and prerequisite relationship between cours.
The information collected from ur through the ur interface is ud to formulate the knowledge graph triples. The triples are finally transformed into RDF statements [15] and stored in xml file using Jena [16]. The xml schema in fig.4 shows cour information in a tree format. This information are transformed into rdf statements. The snippet of the cour knowledge graph automatically constructed in rdf format is shown in fig.5. RDF format is not the only way to reprent knowledge graph. For example, FB15k-237 and NELL-955 knowledge graphs which are ud as datats to evaluate rearch works on knowledge graph completion [17]–[19] are plain text files and the information is not enclod in any tag. However, in this work we ttled for RDF format becau it avail us the opportunity to easily encode relations of interest and other properties (attributes) of entities. Furthermore, manipulation (querying) of the graph is easier.
6Development of Knowledge Graph for University Cours Management
Fig.3. System architecture
Fig.4. xml diagram (schema) showing the nodes in tree format
As stated earlier, the KG is in RDF format. Fig.4 shows one of the rdf triple from the KG and illustrates of how the knowledge graph triples can be deduced. For example, KG triple (CS323, taught_in_2_2018-2019, Dr S.A) – Dr S.A taught CS323 in the cond mester of 2018-2019 can be deduced. Similarly, triple (CS312, prerequisite, CS323) – CS312 is the prerequisite of CS323 can be deduced.
