Design of a Unique Deep Learning Framework for Enhancing High-school Student-inclination Prediction through Bioinspired Computing Model Augmentation
High school students (from classes 9 th to 12 th ) are always in a dilemma about their future prospects. These students usually do not know their strengths & weaknesses, due to lack of exposure, and knowledge about career options. To assist these students, a wide variety of recommender models ar...
Saved in:
| Published in: | 2022 IEEE World Conference on Applied Intelligence and Computing (AIC) pp. 846 - 853 |
|---|---|
| Main Authors: | , |
| Format: | Conference Proceeding |
| Language: | English |
| Published: |
IEEE
17-06-2022
|
| Subjects: | |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | High school students (from classes 9 th to 12 th ) are always in a dilemma about their future prospects. These students usually do not know their strengths & weaknesses, due to lack of exposure, and knowledge about career options. To assist these students, a wide variety of recommender models are designed by researchers, which analyze student's inclination in terms of different career-specific metrics. These metrics include, capability to solve Mathematical tasks, efficiency of reading, ability to identify patterns, etc. But most of these models are general purpose in nature, which limits their performance when applied to specific category of students. To overcome these limitations, this text proposes design of deep learning framework for improving high-school student-inclination prediction via augmented bioinspired computing model has been proposed. This model is able to extract student-specific information via their social media data, family history data, subject wise performance, interest preference, and psychological questionnaire, to train a deep learning model. This model uses VGGNet-19 with existing student behaviour datasets in order to generate a pre-trained classifier which can categorize the student into 1 of N classes via student-specific features. These classes include inclination of student towards engineering, social science, journalism, accounting, and medical fields. It further uses Genetic Algorithm to prioritize different student features based on feature-specific accuracy & precision performance. This model was tested on a large student dataset, and showcased an accuracy of 95.3' for categorizing the student into different inclination classes. |
|---|---|
| DOI: | 10.1109/AIC55036.2022.9848859 |