RESEARCHES
Factors Affecting Mental Health Status of Techworkers
Mental health plays an important role like physical health. As a part of technology advancement lifestyle has become complicated and more stressful. In our research, we detect the elementary reasons which contribute to the most in diminishing mental fitness of people working in technological workplaces.We use ensemble techniques to select distinctive features from our collected data sets. In each stage of workflow, we calibrate the performance of the model applying different classification techniques such as LR, KNN, SVM and CNN. From analysis, we identify the features to be the imperative causes of mental health degradation. More importantly, we have attempted to identify which data mining techniques could be adopted composedly to pave the way to effectively ascertain the features affecting mental health of people providing service at technological workplaces.Moreover from our study, we have observed that the performance on the reduced dataset is satisfactorily analogous to that on actual dataset due to our better feature selection method.
Conference Name: COMPSAC 2020 - IEEE Computer Society Signature Conference on Computers, Software and Applications.
Status: Submitted.
Algorithms for Finding Planted Motif Search in DNA
We have researched on the algorithms for finding special type of motif named (l, d) motif where integer "l" indicates the length of the motif to be discovered and integer "d" indicates the maximum number of mutations (mismatches) allowed in that particular motif. This motif search problem falls in the category of Planted Motif Search (PMS). It takes n strings and two integers "l" and "d" as input. So, our target is to find all possible motifs "M" of length "l" that appear in each of the input sequences with at most "d" mutations. Our proposed heuristic approach is a slight improvement of the previously proposed algorithms of PMS8 and qPMS9. We have improved the neighborhood generation technique by rearranging the DNA sequences in descending order according to their profile matrix value before applying the pruning condition. So, the probability of finding the candidate motif, within few iterations , increases significantly. This will improve the runtime of motif search algorithm to a considerable extent. Our heuristic approach, therefore, betters the runtime of previously worked algorithms.