Vous êtes sur la page 1sur 12

Guru Nanak Dev University, RC

Jalandhar

Sentiment Analysis and Review Classification

Session: 2015 - 2019

SUBMITTED BY:
SUBMITTED TO: Dinesh Pabbi 2015CSA1453
Dr. Pankaj Deep Kaur Kunal Verma 2015CSA1476
Overview and Scope
Review Classification is the task of classifying the reviews given by the users or
customers of the particular service, distinguishing the reviews into different
categories such as a positive review or a negative review.

User Book a Cab for a


Ride

Determination about
amendments needed User used the service
on the basis of reviews.

Gives Feedback /
Feedback is analyzed
Reviews about his/her
by the company
Experience
Aim of the Project

The Aim of the Project is the analyze the feedback / review given by the user in
the textual basis and rating , So that the process of Sentiment analysis can be
carried out in order to determine the various factors that leads to the positive
and negative reviews of the Service.

Collect Reviews
from Online
Forums

Sentiment
Analysis

Fetch
Meaningful
Information
Architecture of the Project

Stop word
Review Data Tokenization
Removal

Text Training Prediction


Vectorization Algorithm Result

 Take the Review Data.


 Tokenize it to break it into Words.
 Remove Stop Words – The Words having no sentiment.
 Vectorize the given data.
 Input the Data Vector into Training Algorithm .
 Generate the Results.
Data Store

Data in .csv
Fed Data to
or json Clean Data
Algorithm
format
Technologies Used for Project

HTML5/CSS/Tkinter JavaScript

Python 3.6 NumPy/Pandas/json/csv

Jupyter
Notebook/Visual Studio
Code 3.6
Jupyter Notebook/ Visual Studio Code
Jupyter Notebook or Visual Studio Code is an Integrated Development Environment that is used for
the development of the project and testing the code.

Python 3.6
The language used to develop the whole project is Python3.6. We Choose this language because It is
one of widely used language for scientific programming.

NumPy
NumPy also known as numerical Python is used to handle the huge amount of data without any
hassle.
Pandas
Pandas is advanced Python library that is used to handle the and wrangle the missing data. It is also
used for cleaning the data before any processing.

HTML5/CSS/Javascript
The above Technologies are used to design the user end of the application.

Keras
The above Technologies are used to design cutting edge deep learning model and train them on GPU
.
Tools Used for the Project
Various Training Algorithms have been implemented to find out the best accuracy on
the trained data.

Logisitc Regression

Support Vector Machine

Recurrent Neural
Network

Convolutional Neural
Network
Machine Learning Algorithms
Challenges

• Overfitting : The training Model can overfit/overlearn according to the specific


data provided so , such model becomes inefficient to generalize.
• Underfitting: The opposite of overfitting , the model over generalizes and hence is
bad for prediction.
• Data Imbalance/Missing Data: To prevent bias, data used for training as well as
testing should be balanced. This also prevents Overfitting to a certain category.
• Language Variance: The language used in different regions of the world is
different and hence it is difficult to build a universal neural network for such a
problem.
About the Institute
Itronix Solutions is one of the most effective Industrial Training Center in the Northern
Region. They are the biggest Corporate Trainer in Mohali and Punjab. Many
Professionals join Itronix Solutions for upgrading and sharpening their technical
skillsets and soft skills. They provide a comprehensive portfolio of Industrial Training
and help learners to achieve and sustain competitive edge in the IT Industry.

We opted Itronix Solutions for our 8th Semester Industrial Training Program because
of the following reasons.

• There is a huge gap between skills required and skills available.


• Many Organizations lack quality manpower especially if it is sourced locally.
• Academic curriculum is rot in line with corporate requirements.
• Huge upfront investment required in training of resources before they start
delivering.
• Itronix Solutions provides best corporate training whether it is updating of technical
skills or soft skills like effective communication, etiquette, better management
practices etc.
References

 https://towardsdatascience.com/naive-bayes-in-machine-learning-
f49cc8f831b4

 https://www.tutorialspoint.com/html/

 https://www.tutorialspoint.com/css

 https://www.tutorialspoint.com/python/python_gui_programming.htm

 https://scikit-learn.org/stable/modules/naive_bayes.html
Thank You.

Vous aimerez peut-être aussi