Author Name:- Anjana Pradeep, Jeena Thomas
Department of Computer Science & Engineering
Abstract:- The aim of this study is to evaluate the performance of students using a fuzzy expert system. The fuzzy process is based solely on the principle of taking non-precise inputs on the factors affecting the performance of students and subjecting them to fuzzy arithmetic to obtain a crisp value of the performance. The system classifies each student's performance by considering various factors using fuzzy logic. Aimed at improving the performance of fuzzy system, several defuzzification methods other than the built methods in MATLAB have been devised in this system for producing more accurate and quantifiable result. This study provides comparison and in depth examination of various defuzzification techniques like Weighted Average Formula (WAF), WAF-max method and Quality Method (QM). A new defuzzification method named as Max-QM which is extended from Quality method that falls within the general framework is also given and commented upon in this study.
Keywords: Fuzzy logic, Fuzzy Expert System, Defuzzification, Weighted Average Formula, Quality Method
I. Introduction
An expert system is a software program that can be used to solve complex reasoning tasks that usually require a (human) expert. In other words, an expert system should help a novice, or partly experienced, problem solver, to match acknowledged experts in the particular domain of problem solving that the system is designed to assist. To be more specific, expert systems are generally conceptualized as shown in Fig 1. The user makes an interaction through the interface system and the system questions the user through the same interface in order to obtain the vital information upon which a decision is to be made. Behind this interface, there are two other sub-systems viz. the knowledge base, which is made up of all the domain-specific knowledge that human experts use when solving that category of problems and the inference engine, a system that performs the necessary reasoning and uses knowledge from the knowledge base in order to come to a judgment with respect to the problem modelled [1].
Expert system has been playing a major role in many disciplines such as in medicines, assist physician in diagnosis of diseases, in agriculture for crop management, insect control, in space technology and in power systems for fault diagnosis[5]. Some expert systems have been developed to replace human experts and to aid humans. The use of an expert system is increasing day by day in today’s world [40]. Expert systems are becoming an integral part of engineering education and even other courses like accounting and management are also accepting them as a better way of teaching[4].Another feature that makes expert system more demanding for students is its ability to adaptively adjust the training for each particular student on the bases of individual students learning pace. This feature can be used more effectively in teaching engineering students. It should be able to monitor student’s progress and make a decision about the next step in training.
Fig. 1: Expert system structure
The few expert systems available in the market present a lot of opportunities for the students who desire more spotlight and time to learn the subjects. Some expert systems present an interactive and friendly environment for students which encourage them to study and adopt a more practical approach towards learning. The expert systems can also act as an assistor or substitute for the teacher. Expert systems focus on each student individually and also keep track of their learning pace. This behavior of an expert system provides autonomous learning procedure for both student and teacher, where teachers act as mentor and students can judge their own performance. Expert system is not only beneficial for the students but also for the teachers which help them guiding students in a better way.
The integration of fuzzy logic with an expert system enhances its capability and is called a fuzzy expert system, as it is useful for solving real world problems which do not require a precise solution. So, there is a need to develop a fuzzy expert system as it can handle imprecise data efficiently and reduces the manual working while enhancing the use of expert system[40].
There are various factors inside and outside college that results in poor quality of academic performance of students[2,3]. To determine all the influencing factors in a single effort is a complex and difficult task. It necessitates a lot of resources and time for an educator to identify all these factors first and then plan the classroom activities and approaches of teaching and learning. It also requires appropriate training, organizational planning and skills to conduct such studies for determining the contributing factors inside and outside college. This process of identification of determinants must be given full attention and priority so that the teachers may be able to develop instructional strategies for making sure that all the students be provided with the opportunities to attain at their fullest potential in learning and performance. By using suitable statistical package it was found that communication, learning facilities, proper guidance and family stress were the factors that affect the student performance. Communication, learning facilities and proper guidance showed a positive impact on student performance and family stress showed a negative impact on student performance. It is indicated that communication is more important factor that affect the student performance than learning facilities and proper guidance [3].
II. Literature review
In this research article seven most important factors are included which affect the students’ performance. These are personal factors, college environment, family factors, and university factors, teaching factors, attendance and marks obtained by students. All these factors are scaled and ranked based on the various sub-factors that are further divided from the base factors. In this study the students’ marks have been focused and not solely on social, economic, and cultural features. To evaluate students’ performance, fuzzy expert system has been developed by considering all the seven factors as inputs to the system. This system has been developed by taking the data of students collected from St. Josephs College of Engineering and Technology, Palai affiliated to M.G University.
II. Literature review
In recent years, many researchers worked on the applications of fuzzy logic and fuzzy sets in educational assessments and grading systems. Biswas[25] presented two methods for evaluating students’ answer scripts using fuzzy sets and a matching function: a fuzzy evaluation method (FEM) and a generalized fuzzy evaluation method. He used fuzzy set theory in student evaluation and found that it is potentially finer than awarding grades or numbers when evaluating answer scripts. He also highlighted that the importance of education system should be to provide students with the evaluation reports regarding their test/examination as sufficient as possible with unavoidable error as small as possible so as to make evaluation system more transparent and fairer to students.
Chen and Lee [26] presented two methods for applying fuzzy sets to overcome the problem of giving two different fuzzy marks to students with the same total score which could arise from Biswas’ method. Their methods perform calculations much faster and complicated matching operations were not required. Echauz and Vachtsevanos [27] proposed a fuzzy logic system for translating traditional scores into letter-grades. Law [28] built a fuzzy structure model with its algorithm to aggregate different test scores in order to produce a single score for individual students in an educational grading system. A method to build the membership functions (MFs) of several linguistic values with different weights was also proposed in this paper.
Wilson, Karr, and Freeman [29] presented an automatic
grading system based on fuzzy rules and genetic algorithms. To assess the
outcomes of student-centered learning using the evaluation of their peers and
lecturer, a new fuzzy set approach was proposed by Ma and Zhou[30] .Wang and
Chen [31] proposed a method for evaluating students answer scripts using fuzzy
numbers associated with degree of confidence. They have considered degree of
confidence of evaluator when awarding satisfaction level to questions of
student answer scripts. Weon and Kim [32] developed an evaluation strategy
based on fuzzy MFs. They pointed out that the system for students’ achievement
evaluation should consider the three important factors of the questions: the
difficulty, the importance, and the complexity. They used singleton functions
to describe the factors of each question reflecting only the individual effect
of the three factors and not the collective effect.
Bai and
Chen [34] pointed out that the difficulty factor is a very subjective parameter
and may cause an argument concerning fairness in evaluation. Bai and Chen [33]
proposed a method for automatically constructing grade membership functions of
fuzzy rules for students’ evaluation. Bai and Chen [34] proposed a method for
applying fuzzy MFs and fuzzy rules for the same purpose. To solve the subjectivity
of the difficulty factor in Weon and Kim’s method [32], they obtained the level
of difficulty as a function of the accuracy of the student’s answer script and
the time consumed to answer the questions. However, their method still has the
subjectivity problem, since the results in scores and ranks are heavily
dependent on the values of several weights which are determined by the
subjective knowledge of domain experts.
In
paper[35], a fuzzy logic evaluation system considering the importance, the difficulty,
and the complexity of questions based on Mamdani’s[36] fuzzy inference and
center of gravity (COG) defuzzification is proposed as an alternative to Bai
and Chen’s method [34]. The transparency and objective nature of the fuzzy
logic system makes it easy to understand and explain the results of evaluation,
and thus to persuade students who are skeptical or not satisfied with the
evaluation results. Li and Chen [37] proposed a method for automatically
generating the weights for several attributes with fuzzy reasoning capability.
Chiang and Lin [38] presented a method for applying the fuzzy set theory to
teaching assessment. Chang and Sun [39] presented a method for fuzzy assessment
of learning performance of junior high school students. From the previous
studies, it can be found that fuzzy numbers, fuzzy sets, fuzzy rules, and fuzzy
logic systems have been used for various educational grading systems.
For more Information CLICK HERE
No comments:
Post a Comment