Browsing by Author "Bethany, Gosala"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Publication
    Detecting design patterns: a hybrid approach based on graph matching and static analysis
    (Springer, 2022) Singh, Jyoti; Chowdhuri, Sripriya Roy; Bethany, Gosala; Gupta, Manjari
    Knowledge and understanding about system design are very important for the development and maintenance of any software system due to certain deadlines and frequent changes in requirements and environment. However, it is a very difficult task to analyse design automatically. Design patterns give standard solutions to common design problems. It is very helpful to find existence of such patterns in the source code. It will reduce effort and time required in understanding and thus in the maintenance activity. In this paper we propose a tool DPDT for detecting design patterns from system software. We use graph matching process to find exact instances of design patterns mapped to system software. In graph matching structural aspects are considered. After that static facts of software systems and design patterns are used to reduce the number of false positives. We evaluate our result on two well-known open source software: JHotDraw and JUnit and compared the result of DPDT with existing tools (Sempatrec, DPF, SSA, DeMIMA, and Depatos) of design patterns detection. It is found that for proxy design patterns our tool out performs the all other tools. Further, for few design patterns it is giving moderate results while other tools did not consider those design patterns. � 2021, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.
  • Thumbnail Image
    Publication
    Machine Learning and Deep Learning Techniques to Classify Depressed Patients from Healthy, by Using Brain Signals from Electroencephalogram (EEG)
    (CRC Press, 2023) Bethany, Gosala; Raj, Gosala Emmanuel; Gupta, Manjari
    Artificial intelligence (AI) has made considerable advancements in the field of Engineering, Science, Technology, and Medicine. Especially, AI techniques like machine learning and deep learning in the field of Medicine are increasing every day, and these techniques will help doctors in the early diagnosis of diseases. Depression is one of the most common and serious mental disorders, which is going to take the first place among mental disorders by 2030. Detecting the depression at early stages needs easy, patient-friendly, and inexpensive method. Through this work, we aim to provide an AI solution for this problem using machine learning and deep learning methods that use EEG signal for the detection of depression. We have extracted 11 statistical features from the signal before feeding them to the classifiers: Logistic Regression, support vector machine, and 1D Convolutional Neural Network. Our methods are tested on a dataset that comprises signals from 30 healthy subjects and 34 MDD patients, and these signals were collected from three different criteria: EC when the eyes of the subject are closed, EO when the eyes of the subject are open, and TASK when the subject is doing some tasks. All three classifiers are applied to each of the three types of signals, which gives a total of nine experiments. Our results found that TASK signals are given better accuracies of 88.4%, 89.3%, and 90.21% for logistic regression, SVM, and 1DCNN, respectively, when compared to EC and EO signals, and our results gave better accuracy than some of the available state-of-the-art methods. � 2024 Rajiv Pandey, Pratibha Maurya and Raymond Chiong.