A Study On Training of Data Set Generation in Machine Learning

Computer Science

The gathering of data is a key component in machine learning and is an ongoing subject of study in many groups. Recently, data gathering has become a major problem for most of two reasons. We are first seeing new applications, which don't necessarily have sufficient labelled data, since master learning is being more utilised. Second, deep learning methods automatically create features that reduce the costs of characteristic procedures, as opposed to conventional machine learning, but, in turn, may need greater quantities of labelled data. Interestingly, current data collector research not only comes from the community of machine learning, natural language and computer vision, but also from the community of data management owing to the significance of huge quantities of information. In this survey, we conduct a thorough examination of data collecting from the point of view of data management. The gathering of data mainly involves acquisition of data, data labelling and the enhancement of existing data or models. Our research landscape provides a guidance on how to utilise these procedures when and identifies intriguing research problems. Integration of mechanical learning and data management in data gathering is part of a major trend towards integration of Big Data with Artificial Intelligence (AI).

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.