A B C D E F G H I J K L M N O P Q R S T U V W X Y Z All
Sheik Abdullah, A.
- Forecasting The Stock Market Values Using Hidden Markov Model
Authors
1 Department of Statistics, Manonmaniam Sundaranar University, Tirunelveli, Tamil Nadu, IN
Source
International Journal of Business Analytics and Intelligence, Vol 4, No 1 (2016), Pagination: 17-21Abstract
The financial market influences personal corporate financial lives and the economic health of a country. Price change of stock market is not a completely random model. The pattern of financial market has been observed by some economists, statisticians and computer scientists. This paper gives a detailed idea about the sequence and state prediction of stock market using Hidden Markov Model and also making inferences regarding stock market trend. The one day difference in close value of stock market value has been used for some period and the corresponding transition probability matrix and emission probability matrix are obtained. Seven optimal hidden states and three sequences are generated using MATLAB and then compared.Keywords
Hidden Markov Model, Transition Probability Matrix, Emission Probability Matrix, Stock Market, States and Sequence.- Comparing the Efficacy of Decision Tree and its Variants using Medical Data
Authors
1 Department of Information Technology, Thiagarajar College of Engineering, Madurai – 625015, Tamil Nadu, IN
2 Department of Computer Science and Engineering, G.K.M. College of Engineering and Technology, Chennai – 600063, Tamil Nadu, IN
3 Department of Medicine, Theni Government Medical College and Hospital, Theni – 625531, Tamil Nadu, IN
Source
Indian Journal of Science and Technology, Vol 10, No 18 (2017), Pagination:Abstract
The objective of this research work focus towards the identification of best variant between decision tree algorithm such as Weighted Decision Trees (WDT), C4.5 Decision Trees and C5.0. Methods: Decision tree has a number of variants such as ID3, Weight based decision tree, C4.5 and C5.0 algorithms. This research work focus towards the predictive performance analysis of weight based decision tree with information gain as the splitting criterion. The algorithm proceeds iteratively with the assignment of weights over the training instances to determine the best among the data attributes. Thereby, the attribute with best weight values can be significantly determined by an improvement over its accuracy. Results: The experimental results proves that among the variants of decision trees the algorithm corresponding to C4.5 provides the highest accuracy of about 71.42% and R2 value of about 0.265 respectively and for real world data the accuracy is about 48.69%. The effectiveness of the decision tree algorithm can be still improved by adopting certain feature selection techniques with the combination of decision tree algorithm. Conclusion: The determined results show that Decision tree algorithm suits well for medical data problems. The efficiency of the algorithm can still be improved by applying Decision Trees for various real world data problems such as Diabetes, Cancer classification with feature selection paradigms. But still a larger set of real world data has to be investigated.Keywords
C4.5 Decision Tree Algorithm, Data Classification, Heart Disease, Predictive Analysis, Weighted Decision Tree,- A Statistical Approach for setting SLO Targets over Outcome Based Education-A Case Study
Authors
1 Department of Information Technology, Thiagarajar College of Engineering, Madurai, IN
Source
Journal of Engineering Education Transformations, Vol 0, No SP 1 (2018), Pagination:Abstract
One of the major challenges in engineering education is to train the students in an efficient way to determine the exposable skills in a collaborative manner to solve complex and weakly structured problems. This research work focus towards the development of a predictive data analytical model using statistical inference techniques. Each of the courses instructed to the students will have certain outcomes to be addressed. Experimental evaluation has been done with the non-applicability of active learning strategies and applicability of active learning strategies over the course outcome values attained by the students over the two batches of students. The data has been formalized to normal distribution to discover the statistical relevance. Newly formulated control limits have been established to determine the course outcome target value. With the initialization of heuristic values we came to a conclusion over the target fixing mechanism for the attainment of course outcomes. Hence from these results we can obviously fix the student target measures for any subjects by which we can deploy the mechanism of deliverables over the reality and the nature of the courses to be learnt.Keywords
Predictive Analysis, Statistics, Education Technology, Control Limits, Course Outcomes, Target Measures.- Impact of Massive Open Online Courses and Best Practices: A Case Study on Social Network Analysis Course
Authors
1 Department of Information Technology, Thiagarajar College of Engineering, Madurai, IN
Source
Journal of Engineering Education Transformations, Vol 31, No 3 (2018), Pagination: 136-140Abstract
Transformations in Engineering Education is to improve the quality of engineering education in learning and research as well as in student development, faculty development, curriculum development and teaching technology methods which involve active learning strategies.Nowadays teaching method takes a new transformation from the conventional method of learning to digital learning like e-learning and m-learning. To improve the students learning ability, a teacher will play a role as mentor/facilitator rather than a teacher. In this paper, we have considered the Impact of Massive Open Online Courses and its Best Practices as a case study on Social Network Analysis Course for the students of third-year Information Technology (IT) department.Keywords
Engineering Education, Massive Open Online Courses, Social Network Analysis, Active Learning Strategies, Course Outcomes.- Assessment of Academic Performance with The E-mental Health Interventions in Virtual Learning Environment Using Machine Learning Techniques: A Hybrid Approach
Authors
1 Department of Information Technology, Thiagarajar College of Engineering, Madurai, Tamil Nadu, IN
Source
Journal of Engineering Education Transformations, Vol 34, No SP ICTIEE (2021), Pagination: 79-85Abstract
Background: The act of virtual learning is defined through learning and practicing in an environment using digital/electronic content for self-paced through online teaching and mentoring. It explicitly deals with the interaction in an asynchronous mode of learning. The quality of teaching-learning depends on the utilization of digital technologies with the advancement in educational technology. There is a need and evaluation for the assessment and estimation of the impact of e-mental health interventions with the students learning through the virtual learning environment.
Purpose/Hypothesis: This research evaluates the psychotherapeutic support for the students to overcome the psychological distress during this COVID-19 pandemic by using machine learning techniques. This mechanism evaluates the efficacy of the academic performance made by the students during the pandemic situation. This analysis involves a hybrid approach for the assessment in machine learning using a genetic algorithm with an artificial neural network upon statistical evaluation. The psychological factors are determined with a keen focus on behaviourism, cognitivism, and social constructivism. The metrics have been evaluated based on digital technologies (ICT) in remote access, individual learning process, flexible learning, cost-effectiveness, time complexity and scalability.
Design/Method: The design process involves the 775 student responses with 27 attributes with differentiation of labels corresponding to behaviourism, cognitivism, and social constructivism. The preprocessed data is fed to genetic algorithm with processing parameters focusing crossover and mutation probability and then classified using artificial neural network. The estimation of academic performance is made using the techniques followed in virtual learning environment such as:
1. Online quiz (Quizizz platform) – Individual assessment
2. Flipped classroom activity - Individual assessment
3. MOOCs online courses – Individual assessment
4. Prototype design – Team activity
5. Research proposal – Team activity
From the assessment process the each of the student performance is evaluated with regard to the course outcome of individual student in the learning environment. The variation has also been observed with the applicability of ALS and traditional practice methods.
Results: The hybrid approach found to be good in the assessment and evaluation of academic performance and health interventions in terms of accuracy (88.18%), precision (94.69%), recall (92.24%), RMS error (0.202) and correlation (0.844) respectively. The statistical analysis and evaluation have been made using Fisher's F-Statistical test, and the P-value is found significantly to be P<0.001. From the experiments, the factors that contribute towards web-based learning, blended learning, and online learning has been differentiated with the psychotherapeutic factors. A total of 775 samples have been used for analysis with the applicability of ICT tools and the pedagogical practices for the course. The factors contributing towards Behaviorism with a focus on interaction and response towards the learning environment plays a significant role in varying the academic performance of the student of about 20% in total learning rate varied significantly. Step-by-step analysis in virtual learning provides a good initiative for the student's community to have a variation in the learning process. Virtual learning is one of the good practices if the ICT in education, process and its principles adhere more efficiently.
Keywords
Virtual Learning, e-Mental Health, Genetic Algorithm, Artificial Neural Network,- A Strategic Approach in Handling Information Retrieval Course for Attaining Course Outcomes – A Case Study
Authors
1 Department of Information Technology, Thiagarajar College of Engineering, Madurai, IN
2 Department of CSE, GKM College of Engineering and Technology, Chennai, IN
Source
Journal of Engineering Education Transformations, Vol 34, No SP ICTIEE (2021), Pagination: 148-153Abstract
Background: The major challenge in engineering education is to educate and train the student’s community with pedagogical practices in order to determine the notable skills to solve more complex and fragile structured problems. Innovative pedagogical practices make the students to adhere to the complex formulations of the corresponding domain and its applications. The realm of good understanding of knowledge and its deliverables can be ascertained through pedagogical practices with a focus on learner-centric activities in the classroom teaching.
The actual target in teaching learning process is to make all the students to have a good exploration on the domain knowledge with success ratio. The learner success is considered to be the core metric with which we can judge the success of learner-centric activities with creativity and quality processes. Confirming the process in which the learners are engaged with the key ideas of the course to be taught makes the students to have practical implementations in Teaching-Learning process.
Objective: This research work focus on the assessment and evaluation of learner-centric techniques for outcome based education upon statistical evaluation. Students of third year (VI Semester) of two consecutive batches have been analyzed for the course on Information Retrieval (14ITPS0). Two set of batches 2015-19 and 2016-20 have been considered for the assessment and analysis of Active Learning Strategies (ALS). The incorporation has been processed using the strategic approach based on daily, weekly and monthly assessments focusing on student learning criteria and their responding behaviors for 2016-20 batch. Each paradigm has been measured corresponding to the course outcomes at each level. Significant statistical analysis has been made for validating the process behind the teaching learning process.
Real time case study: Considering the batch 2015-19 summary assignments, presentations has been used for the evaluation and assessment of interim assignments. The observed response from the students went well among the student’s community. All the students have been registered for the course and they too have completed the same. But, the time and success ratio in learning mechanism varied from student to student upon completion of the course. Considering the batch 2016-20 active learning strategies such as Quiz by Kahoot!, Flipped classroom activity and MOOCs course has been incorporated. Also, for MOOC’S online courses a set of two courses has been identified relevant to the subject of study such as Text Retrieval and Search engines & Text mining and Analytics. For the Research problem identification the students are allowed to choose the domain specific topics for the course on information retrieval. In this process the time and success ratio has also been observed it also varied from one student to another.