无需机器学习、深度学习的文本分类

Question

I want to make a text classifier. But without using any found classification algorithm.Maybe I'll use twitter data for classification. Therefore, I have to classify somehow without training data set. For now I am thinking of using word frequencies for Classification.I can't find any projects that match my thoughts. I need any project or article Can you help me

Answer 1

You can use rule based approaches. For example, you can define some keywords to assign to each class. But it is clear that we can't define all possible keywords in every class. So we solve this problem based on machine learning.

Answer 2

Wow, First off, that is a difficult task and given how well machine learning works usually, for tasks like these. I urge you to try and find training data for your classifier and use machine learning (I find Textblob to be a great and easy to use library for text classification).

To answer your question more directly, you really have to think abstract with this one as they are tons of potential things you can try that yield reliable results. Though Word2Vec works through machine learning, there are lots of interesting and useful concepts within it. See the Wikipedia page here for more details. For example, you can take a look at "word embeddings". Additionally, concepts such as cosine similarity might be of use.

Happy coding!