Distribution Of Training Instances Research Articles

Deep transferable learning on heartbeat classification for imbalance dataset

Electrocardiogram (ECG) data recorded by medical devices are hard to analyze manually. Therefore, it is important to analyze and categorize each heartbeat using machine learning. Recently, advancements in machine learning have made classification of complex data easy and fast. However, these machine learning algorithms require sufficient amount of training data and have limited performance in case the data is imbalance. In case of MIT-BIH arrhythmia dataset, the distribution of training instances are quite imbalance. Many machine learning, particularly deep learning, algorithms give high accuracy on these datasets but still the minority classes have zero accuracy. In this paper, we improve the accuracy of minority classes without hurting the overall accuracy of other classes using transfer learning. The accuracy of existing deep learning model is increased from 90.67% to 98.47%, respectively.

A New Approach for Handling the Iris Data Classification Problem

In this paper, we present a new method to deal with the Iris data classification problem based on the distribution of training instances. First, we find two useful attributes of the Iris data from the training instances that are more suitable to deal with the classification problem. It means that the distribution of the values of these two useful attributes of the three species (i.e., Setosa, Versicolor and Virginica) has less overlapping. Then, we calculate the average attribute values and the standard deviations of these two useful attributes. We also calculate the overlapping areas formed by the values of these two useful attributes between species of the training instances, the average attribute values, and the standard deviations of the values of these two useful attributes of each species. Then, we calculate the difference between the values of these two useful attributes of a testing instance to be classified and the values of these two useful attributes of each species of the training instances. We choose the species that has the smallest difference between the values of these two useful attributes of the testing instance and the values of these two useful attributes of each species of the training instances as the classification result of the testing instance. The proposed method gets a higher average classification accuracy rate than the existing methods.

Data scarcity, robustness and extreme multi-label classification

The goal in extreme multi-label classification (XMC) is to learn a classifier which can assign a small subset of relevant labels to an instance from an extremely large set of target labels. The distribution of training instances among labels in XMC exhibits a long tail, implying that a large fraction of labels have a very small number of positive training instances. Detecting tail-labels, which represent diversity of the label space and account for a large fraction (upto 80%) of all the labels, has been a significant research challenge in XMC. In this work, we pose the tail-label detection task in XMC as robust learning in the presence of worst-case perturbations. This viewpoint is motivated by a key observation that there is a significant change in the distribution of the feature composition of instances of these labels from the training set to test set. For shallow classifiers, our robustness perspective to XMC naturally motivates the well-known ell _1-regularized classification. Contrary to the popular belief that Hamming loss is unsuitable for tail-labels detection in XMC, we show that minimizing (convex upper bound on) Hamming loss with appropriate regularization surpasses many state-of-the-art methods. Furthermore, we also highlight the sub-optimality of the co-ordinate descent based solver in the LibLinear package, which, given its ubiquity, is interesting in its own right. We also investigate the spectral properties of label graphs for providing novel insights towards understanding the conditions governing the performance of Hamming loss based one-vs-rest scheme vis-à-vis label embedding methods.

Text categorization based on genetic fuzzy classification and Boosting method

A novel method for text categorization,which is based on boosting fuzzy classification,was proposed in the paper.Latent Semantic Index(LSI) was used to select text feature and then Boosting algorithm was proposed to integrate fuzzy classification.In each iteration training of boosting algorithm,the distribution of training instances was adjusted,and classification rules were created by genetic algorithm.The weights of the training instances that were classified correctly by available rules were reduced,so that the new fuzzy rule focuses on the misestimate or uncovered instances.Experimental results show that classifier based on fuzzy classification is effective and efficient.