Siddharth-Latthe-07
commented
2 months ago @tisha-prog , i guess you didn't preprocessed the data before performing sentiment analysis through vader, do data cleaning by removing stop words and emojis to get filtered data, consisting of valid comments. After that you can use
SentimentIntensityAnalyzer(), to convert the comments into set of tokens, which would be further sent as an input to VADER.
Refer:- https://github.com/Siddharth-Latthe-07/Sentimental-Analysis-through-Word-Cloud
I have a social media dataset with 500-600 entries, the results of Vader analyzer is driven by emojicon logic ONLY and is ignoring the text in the sentence
Examples -
[gaurav bhai se ye ummid nahi thi 🙂 unsubscribe kr rha hu 🙌🏻 => positive ] The case above the user is unsubscribing to the channel but it is mapped as a positive sentiment
[areee pehle desh me bhukhmari he use hi thik kar lo........rehne ke liye chat nhi he.....rojgar nhi he.......😢😢😢
.
ye badi badi baato se kya hoga jab aam janta be hali se gujar rahi he... => positive] The case above contains the afraid emoji but still, it is a positive sentiment
while the cases below are all negative only because the emojis have a negative sentiment associated with them 1 nhi kr sakte hai 😢
2 that website is not working 😕
3 narender damodardas modi💥
4 कौन कौन मुस्लिम को भारत से बाहर निकालना चाहते हो😢😢
5 i need yogi 😢😢
6 godi channel 🔥🔥
7 dhruv rathe kitnaa agyaani hai jo bhagwan shree ram k existence baare mey nhi jaanta 😢
8 german shepherd ne iske upar video kio nehi banate hei ???😢😅
9 your pm rong😢 not pm
12 this was my project idea at sih hackathon in november, now taken by the government nice 😮😢
13 pls tell us the story of anglo-abor war plsssss 😢😢😢😢😢
14 please tell us about the story of anglo-abor war plsssss 😢😢😢😢😢😢
16 narendra modi 🔥🔥🔥
17 u think soo bro 😢