I want to summarise my document to about 7-8% of total words in document.The summary should be abstractive not extractive. I've referred some of the previous abstractive summarisation strategies however they're using deep learning models like seq2seq, lstm, etc. And I would like to do this task using some basic supervised Machine learning algorithms like svm, logistic regression, etc. The accuracy is not a concern for me. I did a lot of research but was not able to get something relevant.
you can find something similar here https://www.researchgate.net/post/How_to_create_an_abstractive_summary_using_supervised_Machine_learning