Incremental Learning for Classification of Unstructured Data Using Extreme Learning Machine

Abstract

Unstructured data are irregular information with no predefined data model. Streaming data which constantly arrives over time is unstructured, and classifying these data is a tedious task as they lack class labels and get accumulated over time. As the data keeps growing, it becomes difficult to train and create a model from scratch each time. Incremental learning, a self-adaptive algorithm uses the previously learned model information, then learns and accommodates new information from the newly arrived data providing a new model, which avoids the retraining. The incrementally learned knowledge helps to classify the unstructured data. In this paper, we propose a framework CUIL (Classification of Unstructured data using Incremental Learning) which clusters the metadata, assigns a label for each cluster and then creates a model using Extreme Learning Machine (ELM), a feed-forward neural network, incrementally for each batch of data arrived. The proposed framework trains the batches separately, reducing the memory resources, training time significantly and is tested with metadata created for the standard image datasets like MNIST, STL-10, CIFAR-10, Caltech101, and Caltech256. Based on the tabulated results, our proposed work proves to show greater accuracy and efficiency.

Trajectory Clustering and k-NN for Robust Privacy Preserving Spatiotemporal Databases

“…Moreover, the proposed technique cannot handle effectively hidden latent patterns, outliers, missing values, corrupted signals, noise, artifacts, and other disturbances. To this end, deep learning techniques and neural networks (e.g., convolution, recurrent or feed-forward neural networks) [20][21][22][23][24], due to their wide range of applications, are being studied to investigate how and to what extent they can improve the performance of our approach for spatiotemporal databases [25]. The main idea of using deep learning, known as feature learning, is to extract more effective representations of spatiotemporal database records to compute similarity between moving objects and perform clustering of mobile objects with similar motion patterns with higher performance and accuracy, such that the privacy preserving techniques achieve much lower vulnerability close to 1 k .…”

Strategic technological determinant in smart destinations: obtaining an automatic classification of the quality of the destination

“…files that usually contain rows and columns of text with titles. The unstructured data (which accounts for about 90% of the available data) is generally binary data that has no identifiable internal structure (Baars and Kemper, 2008;Madhusudhanan et al, 2018). With the whole data set, we can transform data into value.…”

Incremental Learning with Self-labeling of Incoming High-dimensional Data

“…In real-life scenario, the streaming data that are generated incessantly over time is unstructured, and it is a tedious task to classify these data as they lack the target class labels. Therefore, the study [10] presented the "Classification of Unstructured data using Incremental Learning approach (CUIL)". The proposed method uses the uCLUST algorithm to cluster meta-data to assign the class-labels to each cluster, and later these labeled data are fed to a feed-forward neural network named "Extreme Learning Machine (ELM)" to incrementally assess the new data batches.…”