WOMEN SAFETY USING MACHINE LEARNING

Abstract

The safety‍of‍women‍in‍cities‍is‍still‍a‍serious‍problem‍that‍scares‍them‍a‍lot‍and‍makes‍them‍lose‍their‍freedom‍ and‍confidence‍because‍of‍the‍frequent‍harassment‍that‍has‍been‍happening.‍The‍presented‍research‍proposes‍a‍machine‍ learning‍approach‍that‍examines‍the‍content‍of‍social‍media‍posts‍to‍determine‍the‍places‍where‍it‍is‍unsafe‍for‍women‍to‍ be.‍ Data‍that‍ has‍ been‍ gathered‍ using‍the‍ Tweepy‍library‍is‍ stored‍locally‍ and‍then‍ processed‍for‍the‍ sake‍ of‍ privacy.‍ NLTK‍ removes‍the‍ noise‍from‍the‍text,‍ and‍TextBlob‍ does‍the‍sentiment‍classification‍taking‍the‍ categories‍ of‍ positive,‍ neutral,‍or‍negative.‍The‍locations‍where‍the‍negative‍posts‍are‍coming‍from‍are‍considered‍as‍the‍most‍dangerous‍places‍ on‍the‍map,‍thus‍ showing‍where‍the‍ public‍is‍concerned‍the‍most.‍ The‍ subsequent‍enhancements‍ refer‍to‍live‍ data‍and‍ more‍sophisticated‍AI‍models‍to‍achieve‍better‍precision.0