ClusterClean: a Weak Semi-Supervised Approach for Cleaning Data Labels.
Kyriaki DimitriadouRahul ManghwaniTimothy C. HoadPublished in: IEEE BigData (2019)
Keyphrases
- semi supervised
- training data
- raw data
- data sets
- labeled data
- data analysis
- data sources
- experimental data
- unlabeled data
- data collection
- data processing
- image data
- knowledge discovery
- data points
- input data
- missing data
- synthetic data
- data quality
- data cleaning
- high dimensional data
- statistical analysis
- probability distribution
- prior knowledge
- pairwise
- high quality
- databases