A comparison of dimensionality reduction methods for retrieval of similar objects in simulation data
High-resolution computer simulations produce large volumes of data. As a first step in the analysis of these data, supervised machine learning techniques can be used to retrieve objects similar to a query that the user finds interesting. These objects may be characterized by a large number of features, some of which may be redundant or irrelevant to the similarity retrieval problem. This paper presents a comparison of six dimensionality reduction algorithms on data from a fluid mixing simulation. The objective is to identify methods that efficiently find feature subsets that result in high accuracy rates. Our experimental results with single- and multi-resolution data suggest that standard forward feature selection produces the smallest feature subsets in the shortest time.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 15013665
- Report Number(s):
- UCRL-CONF-200095; TRN: US200803%%1091
- Resource Relation:
- Conference: Presented at: SIAM Conference on Data Mining, Orlando, FL, United States, Apr 22 - Apr 24, 2004
- Country of Publication:
- United States
- Language:
- English
Similar Records
Initial Experiences with Retrieving Similar Objects in Simulation Data
Feature Extraction and Selection From the Perspective of Explosive Detection