I do not know if this is the right place to post my question. I am a beginner in data science and machine learning techniques. I would need to build a modem that can allow me to classify texts and run some sentiment analysis. Right now I have only data collected and they miss any information about sentiment analysis.

Data collected (texts) are approximetely 50000 and they where already cleaned of punctuation and stop words. I heard about the possibility to build some neural networks or use a logistic regression, but I do not know anything about specific models to use for that.

Furthermore, I might consider to build a new model from scratch (I know that can take ages and a lot of efforts) but to do it ai would need to know what I would need (for example, already existing model/dataset to train with dictionaries and sentiment analysis). The problem is that what inwould like to do is to classify a text as positive or negative (sentiment); also, classify it as fake or not fake.

Do you have any suggestion or advice? If you need further information, I would be happy to provide them.


To train a model on your collected data, you need to label your text first. For example, given a sentence

Cats can fly.

It need to be labelled as 'Faked'. Otherwise your model wouldn't know how to classifier between 'Faked' and 'Not Faked'.

However, labelling needs huge effort. So if your data doesn't have any label, I suggest you use a trained model to predict or train a model on an existing sentiment classification dataset. You can search on the Google for the appropriate model of sentiment analysis.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.