Affiliation:
1. Itmo University, St. Petersburg, Russia
2. St. Petersburg Institute for Informatics and Automation of Russian Academy of Sciences (SPIIRAS), St. Petersburg, Russia
Abstract
With the development of the Big Data, data analysis technology has been actively developed, and now it is used in various subject fields. More and more non-computer professional researchers use machine learning algorithms in their work. Unfortunately, datasets can be messy and knowledge cannot be directly extracted, which is why they need preprocessing. Because of the diversity of the algorithms, it is difficult for researchers to find the most suitable algorithm. Most of them choose algorithms through their intuition. The result is often unsatisfactory. Therefore, this article proposes a recommendation system for data processing. This system consists of an ontology subsystem and an estimation subsystem. Ontology technology is used to represent machine learning algorithm taxonomy, and information-theoretic based criteria are used to form recommendations. This system helps users to apply data processing algorithms without specific knowledge from the data science field.
Reference41 articles.
1. Multiple instance classification: Review, taxonomy and comparative study
2. Anastácio, I., Martins, B., & Calado, P. (2011). Supervised Learning for Linking Named Entities to Knowledge Base Entries. In TAC 2011.
3. A web of things approach for indoor position monitoring of elderly and impaired people
4. k-means++: The advantages of careful seeding.;D.Arthur;Proceedings of the eighteenth annual ACM-SIAM symposium on Discrete algorithms,2007
5. Ayodele, T. O. (2010). Types of machine learning algorithms. In New advances in machine learning. IntechOpen.
Cited by
6 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献