International Journal of Research Publication and Reviews, Vol 2, no 12, pp 682-688, December 2021

International Journal of Research Publication and Reviews

Journal homepage: www.ijrpr.com ISSN 2582-7421

Monitoring Suspicious Discussions in online Forums using Data

Mining

Jerusha K Thomas1, Prof. Redhya M2, Dr.T. Mahalekshmi3

1
Final year Student, Sree Narayana Institute of Technology Kollam, Kerala, [email protected]
2
Assistant Professor, Sree Narayana Institute of Technology,Kollam,Kerala, [email protected]
3
Principal, Sree Narayana Institute of Technology,Kollam,Kerala, [email protected]

ABSTRACT

‖IDENTIFY THE SUSPICIOUS DISCUSSION IN COMMON CHAT PLATFORM‖.This system helps to monitor suspicious words and discussions in data
online forms. When a user sends a message system will check the data set with the suspicious data set through data mining. So, if it contains suspicious words
then the user will get blocked. To unblock the user only the admin or moderator can do it.

Keywords— Suspicious word, Spam Detection, Machine Learning.

Introduction

The Project entitled “Monitoring Suspicious Discussions in Online Forms” aimed to identify the suspicious discussions in common chat.
Lately individuals are dependent on the online media like anything, it has turned into the part and bundle of our life. What's more we have begun
utilizing it as a live stage to communicate or sentiments, conclusions, advancements of the recent developments on any point. Extorti on or misinformed
individuals don't pass on any space to spread crimes and web-based media is one of the well known mode of them. Information mining is the strategy
by which we can keep eyes via web-based media. In this system the suspicious word will be detected and block the user. In this system we are used a
spam detection algorithm. The main Objective of this system is
1) To reduce the suspicious Activity on Online forums
2) To notify the Admin about the malicious user
3) To identify and block the user.
In the existing system any fake users/public users can chat any topic without any restriction. The users cannot join a group without admin permission.
The digital technology has been impacting human behavior for a very long time.
In existing system a user cannot enter a group without admin permission or link. The suspicious word is not recorded while chatting. If we want to
block a user then many of the users want to report that user. The existing system analyze text sources from social media and classify the text into
different groups. The system distinguishes between legal and illegal data.
Identify the suspicious discussion in common chat platforms. Our goal is to achieve sentiment analysis for data provided from discussion forums for
which we will build Classifiers which consists of different machine learning classifiers.

EXISTING SYSTEM

How it actually works

In the existing system any fake users/public users can chat any topic without any restriction. The users cannot join a group without admin permission.
The advanced innovation has been affecting human conduct for quite a while.The current System break down text sour ces from online media and
arrange the text into various gatherings. The framework recognizes lawful and illicit information.

Disadvantages of Existing System

 No user can join the group without permission.

 The user has no restrictions for sending a suspicious message.
 The malicious user will be block only after the users report the account.
 International Journal of Research Publication and Reviews, Vol 2, no 12, pp 682-688, December 2021 683

PROPOSED SYSTEM

Identify the suspicious discussion in common chat platforms.

A. Description

Data mining can be utilized to screen web-based media just as conversation gatherings for dubious inputs or remarks. Conversation gatherings can be
utilized to spread any message to an enormous populace quickly. A great many individuals share their perspectives and thought s on legislative issues,
religion and there are additionally individuals who purposefully hurt strict or racial opinions through malignant posts. Henceforth it becomes vital to
screen the posts on these forums. This application gathers the postings and remarks from the conversation locales and investigations those remarks
utilizing information mining procedures and calculation. The gathered information will be analysed for inciting posts utilizing a bunch of watchwords
in the calculation. Further the arrangement of delicate catchphrases are partitioned into 6 classes: - hacking, sexuality, strict, theft, betting,
misrepresentation. In case the remarks in text corpus runs over any of catchphrases identified with any of the 6 classificati ons, then, at that point, it is
ordered into that specific classification to which the watchword has a place.

B. Advantages

• To Reduce the suspicious Activity on Online forums

• To notify the Admin about the malicious user
• To identify and block the user.

Main module of the system

Admin Module:

Approve the moderator registration.In chat contains suspicious words then the user will get blocked. Unblock the user See all the moderators and each
group and group participants.

Moderator Module:

Create groups and admit the public user to the groups. Report the public user when suspicious words identified by system. Unblock the user. See all the
group members in his respected group.

Public User Module:

See all the groups and their moderators

SYSTEM ARCHITECTURE

The architecture of this system mainly consists of three layers: UI Layer, Repository Layer, Application Core Layer.

 UI Layer : It's the most outer layer. It very well may be the web application, Web API, or Unit Test project. This layer has an execution of
the Dependency Inversion Principle with the goal that the application assembles an approximately coupled application. It imparts to the
interior layer by means of interfaces.

 Repository Layer: The layer is expected to make a reflection layer between the Domain elements layer and the Business Logic layer of an
application. It is an information access design that prompts an all the more inexactly coupled way to deal with information access. We make
a nonexclusive vault, which inquiries the information hotspot for the information, maps the information from the information source to a
business element, and endures changes in the business substance to the information source.
 Application Core Layer : It is the middle piece of the design. It holds all application space objects. In case an application is created with
the ORM substance system then this layer holds POCO classes (Code First) or Edmx (Database First) with elements. These space elements
don't have any conditions.
 International Journal of Research Publication and Reviews, Vol 2, no 12, pp 682-688, December 2021 684

VI. Pre-Requisites

A. ASP.NET

Visual studio.net is a finished arrangement of improvement devices for building ASP Web applications, XML Web administrations, work area
applications and versatile applications. Visual basic.NET, Visual C++ .NET all utilization a similar incorporated advancement climate (IDE), which
permits them to share instruments and works with in the making of blended language arrangements. Also, these dialects influence the usefulness of the
.NET Framework, which gives admittance to key advances that improve on the improvement of ASP Web applications and XML Webservices.
ASP.NET joins exceptional engineer usefulness with execution, dependability, and sending. ASP.NET assists with conveying certifiable Web
applications in record time. • Easy Programming Model. ASP.NET makes assembling genuine Web applications drastically more straightforward.
ASP.NET server controls empower a HTML-like style of definitive programming that permits assemble incredible pages with definitely less code than
with exemplary ASP. Showing information, approving client input, and transferring records are generally incredibly simple. The best part is that
ASP.NET pages work in all programs—including Netscape, Opera, AOL, and Web Explorer. Flexible Language Options. Dissimilar to exemplary
ASP, which upheld just deciphered VB Script and J Script, ASP.NET presently support more than 25 .NET languages (including worked in help for
VB.NET, C# and J Script .NET – no instrument required), giving exceptional adaptability in any decision of language.

B. SQL SERVER
SQLServer is really a group of items that consolidate to make up an extremely amazing and rich social information base, stack ed with highlights. The
highlights are: • Relational Database Engine • Replication • Data Transformation Services(DTS) • Analysis Services • Meta Data Services • English
Query The social information base motor is the core of SQL Server and is profoundly adaptable. It likewise gives numerous powerful UIs, they are: •
Enterprise Manager • Query Analyzer SQL Server upholds running various occasions of the social data set motor on a similar PC. Every PC can run
one case of the social data set motor from SQL Server rendition 6.5 or 7.0, alongside at least one example of the data set motor from SQL Server. Each
occasion has its own arrangement of framework and client information bases. Applications can interface with each example on a PC like the manner in
which they associate with occasions of SQL Server running on various PCs. The SQL Server utilities and organization apparatuses have been upgraded
to work with different cases.

C. ALGORITHM USED

Spam Detection Algorithm

Spam detection is a supervised machine learning problem. This implies you should furnish your AI model with a bunch of instances of spam and
 International Journal of Research Publication and Reviews, Vol 2, no 12, pp 682-688, December 2021 685

ham messages and let it track down the pertinent examples that different the two unique classes. We should check out the most effective method to
assemble classifier frameworks that can recognize spammers what's more non spammers. We consider the support vector machine (SVM) and k-
closest neighbor (k-NN) algorithms on the grounds that they address different ways of taking advantage of non content features. While SVMs focus to
the edges and cases close the isolating hyperplanes, k-NN focuses on the positive and negative cases.
Since SVMs are well known classification calculations, we give just a general portrayal of them here; technical subtleties are accessible somewhere
else.
The k-NN calculations are based on the idea of distance between instances. For each test information example, k-NN first finds the top k closest
neighbors as indicated by the distance measure. It then, at that point, observes a weighted majority class among the conceivable class marks. Loads
can be acquainted with reflect distances the test example. When utilized with the non content. highlights, this calculation is simple to execute and can
normally consolidate organization and worldly highlights. The k-NN algorithm is one of the most regularly utilized non content include calculations in
earlier writing, so we can use it as a standard algorithm for examination.

VII. SCOPE

The proposed system is ―Monitoring Suspicious Discussion in Online Forms‖. The project has covered almost all the requirements. Further
requirements and improvements can easily be done since the coding is mainly structured or modular in nature. We can improve this chat web to chat
app. Then we can identify suspicious word from voice note. Another future enhancement is Report abuse. The project has a very vast scope in future.
The project can be implemented on intranet in future.

VIII. RESULT

Fig A.1: Home page

Fig A.2: Login page

International Journal of Research Publication and Reviews, Vol 2, no 12, pp 682-688, December 2021 686

Fig A.3: Registration page

Fig A.4: Admin Page

Fig A.5: User Page

 International Journal of Research Publication and Reviews, Vol 2, no 12, pp 682-688, December 2021 687

Fig A.6 : Moderator Page

Fig A.7 : Blocked Users Page

IX. CONCLUSION

Detection of spam is significant for getting message. Exact identification of spam is a major issue and numerous location strategies have been proposed
by different specialists . Nonetheless, these strategies have an absence of capacity to recognize the spam precisely and proficiently. To address this
issue, we have proposed a strategy for spam location utilizing AI ,that is "Checking dubious conversations in internet based discussions". The primary
target is to screen the dubious movement that happens in different internet based gatherings. This application ful lfills every one of our targets. From the
time the client login and his conversations on any theme accessible in web-based discussion will be checked. When the dubious word is observed then
the client will be block and is advised to executive The "Observing Suspicious Discussion in Online Forms" will really look at the client's message and
rundown (assuming any) dubious words found and will illuminate the administrator.

References

[1] M. Bassiouni, M. Ali, and E. A. El-Dahshan, ―Ham and spam E-mails classifification using machine learning techniques,‖ Journal of Applied
Security Research, vol. 13, no. 3, pp. 315– 331, 2018.

[2] C.J.C. Burges, ―A Tutorial on Support Vector Machines for Pattern Recognition,‖ Data Mining and Knowledge Discovery J., vol. 2, no. 2, 1998,
pp. 121–167.

[3] Artificial Intelligence & Machine Learning ,Vinod Chandra & Hareendran A.
 International Journal of Research Publication and Reviews, Vol 2, no 12, pp 682-688, December 2021 688

[4] Jiwan Han & Micheline Kamber – Data Mining Concepts & Techniques – 2ed-.Elsevier.

[5] F. Sebastiani, ―Machine Learning in Automated Text Categorization,‖ ACM Computing Surveys, vol. 34, no. 1, 2002, pp. 1–47.

[6] https://www.c-sharpcorner.com/article/onion archi tecture-in-asp-net-core-mvc/

[7]https://www.researchgate.net/publication/336252196_MONITORING_OF_SUSPICIOUS_DISCUSSIONS_ON_ONLINE_FORUMS_USING_DA
TA_MINING

[8]https://docs.microsoft.com/enus/aspnet/core/mvc/overview?view=aspnetcore-6.0