LEARNING TO PREDICT CLOSED QUESTIONS ON STACK OVERFLOW

G. Lezina, A. Kuznetsov, P. Braslavski

Research output: Contribution to journalArticle

Abstract

The paper deals with the problem of predicting whether the user's question will be closed by the moderator on Stack Overflow, a popular question answering service devoted to software programming. The task along with data and evaluation metrics was offered as an open machine learning competition on Kaggle platform. To solve this problem, we employed a wide range of classification features related to users, their interactions, and post content. Classification was carried out using several machine learning methods. According to the results of the experiment, the most important features are characteristics of the user and topical features of the question. The best results were obtained using Vowpal Wabbit - an implementation of online learning based on stochastic gradient descent. Our results are among the best ones in overall ranking, although they were obtained after the official competition was over.
Original languageEnglish
Pages (from-to)118-133
Number of pages16
JournalУченые записки Казанского университета. Серия: Физико-математические науки
Volume155
Issue number4
Publication statusPublished - 2013

GRNTI

  • 28.23.00

Level of Research Output

  • VAK List

Fingerprint Dive into the research topics of 'LEARNING TO PREDICT CLOSED QUESTIONS ON STACK OVERFLOW'. Together they form a unique fingerprint.

Cite this