1
|
Wang Y, Li Y, Qiao C, Liu X, Hao M, Shugart YY, Xiong M, Jin L. Nuclear Norm Clustering: a promising alternative method for clustering tasks. Sci Rep 2018; 8:10873. [PMID: 30022093 PMCID: PMC6052164 DOI: 10.1038/s41598-018-29246-4] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/26/2017] [Accepted: 07/02/2018] [Indexed: 11/09/2022] Open
Abstract
Clustering techniques are widely used in many applications. The goal of clustering is to identify patterns or groups of similar objects within a dataset of interest. However, many cluster methods are neither robust nor sensitive to noises and outliers in real data. In this paper, we present Nuclear Norm Clustering (NNC, available at https://sourceforge.net/projects/nnc/), an algorithm that can be used in various fields as a promising alternative to the k-means clustering method. The NNC algorithm requires users to provide a data matrix M and a desired number of cluster K. We employed simulated annealing techniques to choose an optimal label vector that minimizes nuclear norm of the pooled within cluster residual matrix. To evaluate the performance of the NNC algorithm, we compared the performance of both 15 public datasets and 2 genome-wide association studies (GWAS) on psoriasis, comparing our method with other classic methods. The results indicate that NNC method has a competitive performance in terms of F-score on 15 benchmarked public datasets and 2 psoriasis GWAS datasets. So NNC is a promising alternative method for clustering tasks.
Collapse
Affiliation(s)
- Yi Wang
- Ministry of Education Key Laboratory of Contemporary Anthropology, Department of Anthropology and Human Genetics, School of Life Sciences, Fudan University, Shanghai, China.,Human Phenome Institute, Fudan University, Shanghai, China
| | - Yi Li
- Ministry of Education Key Laboratory of Contemporary Anthropology, Department of Anthropology and Human Genetics, School of Life Sciences, Fudan University, Shanghai, China.,Six Industrial Research Institute, Fudan University, Shanghai, China.,Human Phenome Institute, Fudan University, Shanghai, China
| | - Chunhong Qiao
- Ministry of Education Key Laboratory of Contemporary Anthropology, Department of Anthropology and Human Genetics, School of Life Sciences, Fudan University, Shanghai, China.,Human Phenome Institute, Fudan University, Shanghai, China
| | - Xiaoyu Liu
- Ministry of Education Key Laboratory of Contemporary Anthropology, Department of Anthropology and Human Genetics, School of Life Sciences, Fudan University, Shanghai, China.,Human Phenome Institute, Fudan University, Shanghai, China
| | - Meng Hao
- Ministry of Education Key Laboratory of Contemporary Anthropology, Department of Anthropology and Human Genetics, School of Life Sciences, Fudan University, Shanghai, China.,Human Phenome Institute, Fudan University, Shanghai, China
| | - Yin Yao Shugart
- State Key Laboratory of Genetic Engineering, Collaborative Innovation Center for Genetics and Development, School of Life Sciences, Fudan University, Shanghai, China. .,Unit on Statistical Genomics, Division of Intramural Division Programs, National, Institute of Mental Health, National Institutes of Health, Bethesda, MD, USA. .,Six Industrial Research Institute, Fudan University, Shanghai, China.
| | - Momiao Xiong
- Human Genetics Center, School of Public Health, University of Texas Houston Health Sciences Center, Houston, Texas, USA.
| | - Li Jin
- State Key Laboratory of Genetic Engineering, Collaborative Innovation Center for Genetics and Development, School of Life Sciences, Fudan University, Shanghai, China. .,Six Industrial Research Institute, Fudan University, Shanghai, China. .,Human Phenome Institute, Fudan University, Shanghai, China.
| |
Collapse
|
2
|
Random Bits Forest: a Strong Classifier/Regressor for Big Data. Sci Rep 2016; 6:30086. [PMID: 27444562 PMCID: PMC4957112 DOI: 10.1038/srep30086] [Citation(s) in RCA: 10] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/02/2016] [Accepted: 06/28/2016] [Indexed: 11/08/2022] Open
Abstract
Efficiency, memory consumption, and robustness are common problems with many popular methods for data analysis. As a solution, we present Random Bits Forest (RBF), a classification and regression algorithm that integrates neural networks (for depth), boosting (for width), and random forests (for prediction accuracy). Through a gradient boosting scheme, it first generates and selects ~10,000 small, 3-layer random neural networks. These networks are then fed into a modified random forest algorithm to obtain predictions. Testing with datasets from the UCI (University of California, Irvine) Machine Learning Repository shows that RBF outperforms other popular methods in both accuracy and robustness, especially with large datasets (N > 1000). The algorithm also performed highly in testing with an independent data set, a real psoriasis genome-wide association study (GWAS).
Collapse
|