FeatureHub the Answer to Big Data Problems

Big data analysis is tricky even for the best of the best. The point of these issues lies in the understanding of the “features”. But the analysis of the features depends upon the human mind and therein lays the problem.
But MIT researchers have found a way out. The researchers developed a new tool called “FeatureHub”. This tool is the easy answer to all big data-related problems and queries. The tool provides the scientists and researchers an efficient and elaborative method of calculating their data, thereby facilitating easy use of the data obtained from big data.
The scientists and researchers can log onto the central website of the tool and put in their problems. Then, they can spend a couple of hours reviewing their problem and bouncing some ideas about the problem. The most amazing feature of the project is that this tool works at understanding the topic. After the topic is put in, the tool looks for various options and combinations of features. After reviewing combinations, the tool gives you the best possible outcome for your problem in your targeted data.
The study had thirty-two analysts participate in the research. The analysts were given some time with the tool to get familiar. Then the scientists spent about five hours each trying to find outcomes for their problems in the targeted area. The results of this were published in a completion where it was compared against another tool called Kaggle.
The research found that the scientists who used Kaggle scored about hundred points. And the scientists who used FeatureHub scored about five to six points lesser than the winners. The best part was that the winning ideas were a work of four to five months of intense hard work and research, whereas FeatureHub helped the scientists produce the outcome in a matter of few days.
FeatureHub’s name is inspired from “GitHub”. “I think that the concept of massive and open data science can be really leveraged for areas where there’s a strong social impact but not necessarily a single profit-making or government organization that is coordinating responses,” says the lead author of the paper, Micah Smith.
The user interface of the tool is made with the help of the suite Jupyter Notebook, and the feature sets are normal software learning packages. The features in the tool must be written in the phyton code, a computer language, and the syntax of the program has been deliberately kept simple.

Leave a Reply

Your email address will not be published. Required fields are marked *