Warm Start for Parameter Selection of Linear Classifiers
345 views · Published 28 October 2015 · 20:00 · Indexed 20 September 2026
Channel: Association for Computing Machinery (ACM) · 2015 · Science & Technology
Authors: Bo-Yu Chu, Chia-Hua Ho, Cheng-Hao Tsai, Chieh-Yen Lin, Chih-Jen Lin Abstract: In linear classification, a regularization term effectively remedies the overfitting problem, but selecting a good regularization parameter is usually time consuming. We consider cross validation for the selection process, so several optimization problems under different parameters must be solved. Our aim is to devise effective warm-start strategies to efficiently solve this sequence of optimization problems. We detailedly investigate the relationship between optimal solutions of logistic regression/linear SVM and regularization parameters. Based on the analysis, we develop an efficient tool to automatically find a suitable parameter for users with no related background knowledge. ACM DL: http://dl.acm.org/citation.cfm?id=2783332 DOI: http://dx.doi.org/10.1145/2783258.2783332
More from this channel
-
0:30
Challenges and opportunities for technology in foreign language classrooms
-
0:31
LongPad
-
0:30
The effects of tactile feedback and movement alteration on interaction and awareness with digital...
-
0:31
Understanding the privacy-personalization dilemma for web search
-
0:33
KINECT wheels: wheelchair-accessible motion-based game interaction
-
0:31
SIGCHI extended abstracts: gravity of light
-
0:30
Complementing Text Entry Evaluations with a Composition Task
-
0:31
Performativity in Sustainable Interaction: The Case of Seasonal Grocery Shopping in EcoFriends