An Efficient Model In Online Education Data Analysis Using Accelerated Algorithm

Document Type : Primary Research paper


1 PG Student, Department of Coputer Science and Engineering, Mahendra Engineering College, Namakkal

2 Associate Professor, Department of Coputer Science and Engineering, Mahendra Engineering College, Namakkal


The main idea of our approach is to analyze sentiment analysis is an issue that
can be solved by using learning management systems to enhance learning via social media.
Existing sentiment categorization has long been recognised as a domain-specific issue. As
a result, it produces a very low estimation depression rate of a person and accuracy is low.
The problem can be rectified by using the classifier that is collaborative multi domain
sentiment classifier, advantage of this is to find the depression state of a person more
accurately. To do this, we will utilise multi-task learning to collaboratively train sentiment
classifiers for different domains in a collaborative way. When it comes to brand messaging,
political campaigns, marketing research, and customer feedback, it is often utilised. The
analysis of the subject The use of a Bag of word is important technique (BODW). The
essential words in a text reflect either a fact or a feeling about the subject. The objective
labels are represented by fact, whereas the subjective labels are represented by emotion.
Based on objective and subjective variable selection, the system extracts a bag of
discriminative words from a text. The discriminative words in a text are filtered using LDA
and regression methods. The SA is also done by one of those methods named support
vector machine, and the classification algorithm of Nave Bayes without the need for
language resources, it is possible to categorise emotion words into positive and negative
categories while simultaneously recording changes. This result will tell us who is in
depression by calculating the score. This can be done with the help of government approval
to access the data from socialmedia.