every week I get a batch of customer reviews. I already created a finetuned Sentiment Analysis Model and a finetuned classification model to evaluate if a review is positive/negative/neutral and to which of 6 distinct topic the review belongs. The reviews are on average pretty short (mean of 11 tokens) and the weekly batch is about 500 reviews. Now my question: Does it make sense to use a Summarization model to get the key phrases for e.g all negative reviews which belong to a specific topic (problems with website). Or are my tweets too short and not enough? Has anyone got any experiences?