A HYBRID LEXICON-BASED MACHINE LEARNING FRAMEWORK FOR POLITICAL SECURITY THREAT PREDICTION

Abstract

The internet is pivotal to public safety
today. The US Knowledge People group positions
digital dangers close by illegal intimidation and other
key difficulties. Safeguarding a country is more
diligently now. With such a lot of information, data on
the online, and fake news, there's a never-ending
chance of impelling scorn and undermining public
safety. Project joins emotions and perspectives to
country dangers. Negative emotions in internet data
could undermine public safety. Specialists should
rapidly recognize and address these opinions.
Emotions and public safety concerns are connected,
but there is little assessment and structure
investigation of feelings and their measurements.
Review have for the most part characterized human
emotions , not their significance to public safety risks
or how to figure their ascent. This idea proposes
foreseeing political risks utilizing web news feelings.
It centers around political security, an essential public
safety issue. The examination utilizes word
investigation and ML to connect information holes and
show their adequacy utilizing genuine web news
information. The exploration utilizes complex troupe
picking up utilizing a stacking classifier, Random
Forest-Decision Tree models, and a standalone
Random Forest model. This group method further
develops framework expectation, conveying
intimidation forecast more strong. Client testing is
empowered through an easy to use Flask framework
with SQLite network and basic information exchange
and signin. It increments model execution and gives a
sensible stage to genuine client collaborations, giving
a total framework ease of use and viability
assessment.