FINAL RESEARCH WORK

Progressive Learning - Universal Classifier

In this project, the progressive learning technique will be integrated with the universal classifier to achieve human-learning-inspired progressively learning universally generic classifier.  The resulting new classifier can be used for binary, multi-class and multi-label classification problems with dynamic class introductions. The newly developed classifier having developed based on the extreme learning machine, exploits its inherent high speed training and testing.

EARLIER RESEARCH WORKS

Progressive Learning - Multi-label Classifier

In this project, an online multi-label classifier integrated with progressive learning technique (PLT) is developed.  The resulting progressive learning technique based online classifier can learn new labels while retaining the knowledge of previous labels for multi-label problems. It is to be highlighted that there are no similar algorithms available in the literature for multi-label classification. The developed technique is evaluated with different multi-label datasets for consistency, performance and speed. 

Progressive Learning - Multi-class Classifier

In this project, the concept of progressive learning technique (PLT) is introduced and demonstrated on multi-class classification problems. The key objective of PLT is that it can dynamically learn new classes on the run. Being a newly proposed technique, the consistency and the complexity of the PLT are analyzed. Several datasets are used to evaluate the performance and speed of the developed technique.

Online Learning - Universal Classifier

In this project, a universal classifier which is capable of classifying binary, multi-class and multi-label classification is developed. There are no such universal classifiers available in the literature thus far. The developed classifier is experimented with datasets from binary, multi-class and multi-label problems and its performance is compared with state-of-the-art techniques from corresponding type of classification problem.

Online Learning - Multi-label Classifier

In this project, a high-speed online neural network classifier based on extreme learning machines for multi-label classification is developed. The proposed work exploits the high-speed nature of the extreme learning machines to achieve real-time multi-label classification of streaming data. A new threshold-based online sequential learning algorithm is proposed for high speed and streaming data classification of multi-label problems. The developed method is experimented with six datasets from different application domains such as multimedia, text, and biology and results are compared with nine different state-of-the-art methods.

Batch Learning - Multi-label Classifier

In this project, a high speed neural network classifier based on extreme learning machines for multi-label classification problem is developed. Multi-label classification is a superset of traditional binary and multi-class classification problems. As opposed to the single-label problem, both the number of labels the sample belongs to, and each of those target labels are to be identified for multi-label classification resulting in increased complexity. The proposed method is applied to 6 benchmark datasets of different domains and a wide range of label density and label cardinality. The results are compared with 9 state-of-the-arts multi-label classifiers.