You all must have seen a post with lots of tweets!!
Every second a new post is share by an individual and each post has lots of comments, Tweets. A post contains lots of thoughts, so its difficult to know how a tweet is...We all give our thoughts on a post so it can be Bad, Good or sometime Worst. So how can we know how many tweets are Positive or Negative.
To analyze the sentiments We are going to extract the words and phrases from the tweets so we can know weather a sentiment Positive, Negative or Neutral. We extracted a dataset which is provided by thr kaggle and we stated the analysis we trained the bot to find the phrase which tells the sentiment of the sentace.
In this project we used bert and roberta model to ectract the phrases that causes sentiment, We can use word_clouds too!! what are word_clouds? An image composed of words used in a particular text or subject, in which the size of each word indicates its frequency or importance.
We used NLP a Natural Language Processing Made Easy - using SpaCy ( in Python). It is one of the principal areas of Artificial Intelligence. NLP plays a critical role in many intelligent applications and is a capacious field, some of the tasks in nlp are – text classification.
roBERTa model which is also one of the bert based model is trained in 5folds: -
- cross validation
- Inference means loading weights of previously trained models
- Predicting on the test set
- Training on the data.
- More Accurate results.
1. Training data: Training data is an initial set of data used to help a program understand how to apply technologies like neural networks to learn and produce sophisticated results.
2. Folds: Also termed reduce, accumulate, aggregate, compress, or inject refers to a family of higher-order functions that analyze a recursive data structure.
3. Cleaning data: Cleaning is the process of detecting and correcting(or removing) corrupt or inaccurate records from a record set, table, or database.
Reference Blog: https://datepranay.wixsite.com/techanalysis/post/best-software-for-analyzing-sentiments