Click Here https://spam-detector-ankur.herokuapp.com/ to redirect to website where it is hosted (not working as free tier over)
Approach - I have data of some text msg . Some are spam and some are not . After cleaning and other preprocessing , i saparate out spam and non - spam msg . Then we will extract words that are common . Actually probability of being spam and non spam is calculated over this
Concept - As stated concept used is **naive bayes algorithm ** in Simple way
P(spam) = no. of spam / total
= 0.13
This mean 13% of total mail generally found as spam
P(text| spam)- all logic lies here
dictionary with key = word and value = probability of occurance is created for both spamer and non- spammer word bag
we will check each work of text entered by user and multiply all to determine spam-probability and non spam probability . Then compare and provide result .
Example = " he is good spammer "
we will multiply probability value ; Both spammer and non -spamer
then comapre