In a data analytic view, informaticists are familiar with unsupervised clustering analysis and class-label supervised clustering analysis.
A 'predicting analysis of microarrays' (PAM) was performed to find genes that accurately predict classes based on class labels (supervised analysis) (Tibshirani et al, 2002), using all 18 336 genes of the array.
By leveraging the labeled information via supervised learning, the learned features are endowed with more discriminative power.
The lack of labeled data for supervised classifier design, user anonymity, and the size of the data-sets are the other key factors differentiating the problem from previous studies, and are the key drivers behind the design and implementation decisions for the system described.
In other words, given labeled training data (supervised learning), the algorithm outputs an optimal hyperplane which categorizes new examples.
The supervised techniques require labeled data, while semi-supervised techniques need guidance from domain experts.
The supervised techniques require labeled data, while the semi supervised techniques need manual tuning from domain experts.
Recently, saliency detection has become an active research topic in learning from labeled image, where various supervised methods were designed.
Therefore, we investigated whether a relatively small amount of labeled data in the supervised learning task would result in better separation compared to unsupervised clustering with our feature set.
Supervised and semi-supervised methods are labeled 'SVM' followed by the percentage of labeled data (10, 30, 50, 70, 90, 100%).
The system is based on the distant supervision paradigm, i.e., using a database, in this case of drugs and their indications and side-effects, to label examples for supervised machine learning.
