Prognosticate lung diseases have becoming a challenge in front of health department. At time of novel corona period lung was the highest affected part of human body. Few of related research have been done but that research has some issues, so in state-of-the-art of that we need to explore it more to early diagnose lung disease. The objective of this paper is, first to find out the lack of literature, second to explore the state-of-the-art of machine and deep-learning techniques used in literature and third analysis their results. In existing research the main issue is overfitting and underfitting problem. A common taxonomy of fivecharacteristicswerecast-off in almost all reviewed articles: image/text data type, data enhancement, extraction of features, sort of classification techniques and kind of lung ailment. Using present taxonomy of research or using some advancement in existing, Experts can increase the reliability of their automated learning framework to progress their research.