Effective threat intelligence is quite significant, as evidenced by the current trend of attacks in cyber security on businesses, growing rapidly and intensively. With that, this paper compares the results of various approaches using a novel multi-layered threat intelligence architecture that combines data from various sources, that include, intelligence from open sources,network logs from corporate hubs, and surveillance of dark web, with a few machine learning(ML) algorithms, like, Decision Tree, Naive Bayes,Linear and Random and Quadratic Discriminant Analysis, Gradient Boosting, Extra Trees,. Ada and XGBoost DNNLST.