Download as pdf or txt
Download as pdf or txt
You are on page 1of 1

Embedding GloVe into the corpus using a multi-layer

BiLSTM deep learning model for opinion mining through


Daraz product reviews
MD. Redwan Hossan, Mohammad Borhan Uddin, Al Nur Istiak, Moriom Sultana

Advisor: Dr. Mohammad Rashedur Rahman

1 Introduction 4 UI Of The System

The objective of this project is to analyze public


sentiment on product review from Ecommerce
website, Daraz. For this project, we built our
dataset by scrapping reviews on various
product of certain brands.

This analysis system has been developed with


the help of Python 3 and Google Colab Engine.

This analysis system would give better


accuracy than most other methods that exist.

2 System Diagram
This is a visual presentation of what user
sees in the project

5 Results

This figure shows the diagram of the system.

3 Model Architecture
It's a synopsis of our recommended remedies and
their outcomes. The procedure is continued until a
predefined meeting model is met, a predetermined
number of eras has passed, or the arrangement has
remained the same for several times.

6 Conclusions

❖ We offer a technique for analyzing sentiment on brands


product automatically. Using a vast dataset of product
reviews, we were able to adequately predict whether a
Bidirectional recurrent neural networks(RNN) are review is POSITIVE or NEGATIVE in a matter of
really just putting two independent RNNs together. seconds using a basic Bi-LSTM structure with an
Using bidirectional will run inputs in two ways, one accuracy of 93%.
from past to future and one from future to past and ❖ Our system can obtain competitive performance in this
what differs this approach from unidirectional is challenge.
that in the LSTM that runs backward you preserve ❖ We intend to investigate ways to improve the resilience
information from the future and using the two
hidden states combined you are able in any point in of our system against sarcasm in reviews utilizing
time to preserve information from both past and unseen approaches during training in the future.
future.

You might also like