Page 7 - Campus Chronicles Technical Magazine 2020

Thus the developed system as well within the MODULES:
TECHNEWIL budget and this was achieved because most of There are five modules divided in this project in

CLUB the technologies used are freely available. order develop the concept of sentiment
Weakly - Supervised ACTIVITIES Only the customized products had to be analysis with tagging. They are listed below

Deep Embedding For purchased. 1. Products Initiation

2. Products acquisition
TECHNICAL FEASIBILITY: This study is carried
Product Review out to check. Any system developed must not 3. Sentiment classification
4. Weak Supervision
Sentiment Analysis have a high demand on the available technical 5. Graphical Analysis
resources. This will lead to high demands on
the available technical resources. This will lead
By Ms. Pranitha Mettu, to high demands being placed on the client. PRODUCTS INITIATION:
The First phase of the implementation of this
B. Tech (ECE) - 20UP1A0428 The developed system must have a modest project is Products Initiation. In this module

requirement, as only minimal or null changes
admin is uploading the products which user
INTRODUCTION: People are getting used to PROPOSED SYSTEM: In this work, we propose are required for implementing this system. wants to see and purchase. Once admin
consuming online and writing comments about a novel deep learning framework for review SOCIAL FEASIBILITY: The aspect of study is to uploads the product means it stored in the
their purchase experiences on merchant/ sentence sentiment classification. The check the level of acceptance of the system by database. The products which are uploaded

review Websites. framework treats review ratings as weak labels the user. This includes the process of training are listed in website to admin in order to modify

These opinionated contents are valuable to train deep neural networks. For example, the user to use the system efficiently. The user or delete the particular product. Admin is the
resources both to future customers for with 5-stars scale we can deem ratings must not feel threatened by the system, only authorized person to upload the products
decision-making and to merchants for above/below 3-stars as positive/ negative instead must accept it as a necessity. The level in this project.

improving their products and/or service. weak labels respectively. The of acceptance by the users solely depends on PRODUCTS ACQUISITION: The second

However, as the volume of reviews grows Framework generally consists of two steps. In the methods that are employed to educate the module of this product conveys that user can
user about the system and to make him familiar
rapidly, people have to face a severe the first step, rather than predicting sentiment with it. His level of confidence must be raised view the products which are uploaded by
information overload problem. labels directly, we try to learn an embedding admin. Then they can view the ratings and
space (a high level layer in the neural network) so that he is also able to make some reviews of the same products which are given
EXISTING SYSTEM: Lexicon-based methods constructive criticism, which is welcomed, as
which reflects the general sentiment by other users who already purchased the
typically take the tack of first constructing a he is the final user of the system.
distribution of sentences, from a large number product. According to the help of ratings and
sentiment lexicon of opinion words (e.g. of weakly labeled sentences. ARCHITECTURE: reviews user can purchase the product. The
“wonderful”, “disgusting”), and then design ordered list is also shown in the project for the
classification rules based on appeared opinion SYSTEM STUDY: Three key considerations convenience of users. The cart and checkout

words and prior syntactic knowledge. Despite 1. Economical Feasibility facility is also available to users from this
effectiveness, this kind of methods requires 2. Technical Feasibility module.
substantial efforts in lexicon construction and 3. Social Feasibility
rule design. Furthermore, lexicon-based ECONOMICAL FEASIBILITY: This study is SENTIMENT CLASSIFICATION: The users who

methods cannot well handle implicit opinions, carried out to check the economic impact that are all purchased the products can rate product
i.e. objective statements such as “I bought the the system will have on the organization. The as per their interest on one scale of five and
mattress a week ago, and a valley appeared amount of fund that the company can pour into they are free to comment for the same. Based
today”. As pointed out in this is also an the research and development of the system is on the ratings and reviews given by user

important form of opinions. Factual information limited. The expenditures must be justified. sentiment can be analyzed. There are two
is usually more helpful than subjective feelings. sentiments maintained in this project they are
Lexicon-based methods can only deal with positive and negative.
Campus Campus
implicit opinions in an ad-hoc way.
CHRONICLES 6 7 CHRONICLES
Technical Magazine Technical Magazine