ENSEMBLE META CLASSIFIER WITH SAMPLING AND FEATURE SELECTION FOR DATA WITH IMBALANCE MULTICLASS PROBLEM

Ensemble learning by combining several single classifiers or another ensemble classifier is one of the procedures to solve the imbalance problem in multiclass data. However, this approach still faces the question of how the ensemble methods obtain their higher performance. In this paper, an investigation was carried out on the design of the meta classifier ensemble with sampling and feature selection for multiclass imbalanced data. The specific objectives were: 1) to improve the ensemble classifier through data-level approach (sampling and feature selection); 2) to perform experiments on sampling, feature selection, and ensemble classifier model; and 3 ) to evaluate t he performance of the ensemble classifier. To fulfil the objectives, a preliminary data collection of Malaysian plants’ leaf images was prepared and experimented, and the results were compared. The ensemble design was also tested with three other high imbalance ratio benchmark data. It was found that the design using sampling, feature selection, and ensemble classifier method via AdaboostM1 with random forest (also an ensemble classifier) provided improved performance throughout the investigation. The result of this study is important to the on-going problem of multiclass imbalance where specific structure and its performance can be improved in terms of processing time and accuracy.

Neural Legal Outcome Prediction with Partial Least Squares Compression

Predicting the outcome of a case from a set of factual data is a common goal in legal knowledge discovery. In practice, solving this task is most of the time difficult due to the scarcity of labeled datasets. Additionally, processing long documents often leads to sparse data, which adds another layer of complexity. This paper presents a study focused on the french decisions of the European Court of Human Rights (ECtHR) for which we build various classification tasks. These tasks consist first of all in the prediction of the potential violation of an article of the convention, using extracted facts. A multiclass problem is also created, with the objective of determining whether an article is relevant to plead given some circumstances. We solve these tasks by comparing simple linear models to an attention-based neural network. We also take advantage of a modified partial least squares algorithm that we integrate in the aforementioned models, capable of effectively dealing with classification problems and scale with sparse inputs coming from natural language tasks.

EEG-Controlled Wall-Crawling Cleaning Robot Using SSVEP-Based Brain-Computer Interface

The assistive, adaptive, and rehabilitative applications of EEG-based robot control and navigation are undergoing a major transformation in dimension as well as scope. Under the background of artificial intelligence, medical and nonmedical robots have rapidly developed and have gradually been applied to enhance the quality of people’s lives. We focus on connecting the brain with a mobile home robot by translating brain signals to computer commands to build a brain-computer interface that may offer the promise of greatly enhancing the quality of life of disabled and able-bodied people by considerably improving their autonomy, mobility, and abilities. Several types of robots have been controlled using BCI systems to complete real-time simple and/or complicated tasks with high performances. In this paper, a new EEG-based intelligent teleoperation system was designed for a mobile wall-crawling cleaning robot. This robot uses crawler type instead of the traditional wheel type to be used for window or floor cleaning. For EEG-based system controlling the robot position to climb the wall and complete the tasks of cleaning, we extracted steady state visually evoked potential (SSVEP) from the collected electroencephalography (EEG) signal. The visual stimulation interface in the proposed SSVEP-based BCI was composed of four flicker pieces with different frequencies (e.g., 6 Hz, 7.5 Hz, 8.57 Hz, and 10 Hz). Seven subjects were able to smoothly control the movement directions of the cleaning robot by looking at the corresponding flicker using their brain activity. To solve the multiclass problem, thereby achieving the purpose of cleaning the wall within a short period, the canonical correlation analysis (CCA) classification algorithm had been used. Offline and online experiments were held to analyze/classify EEG signals and use them as real-time commands. The proposed system was efficient in the classification and control phases with an obtained accuracy of 89.92% and had an efficient response speed and timing with a bit rate of 22.23 bits/min. These results suggested that the proposed EEG-based clean robot system is promising for smart home control in terms of completing the tasks of cleaning the walls with efficiency, safety, and robustness.

Off-Topic Essay Detection: A comparative study on the Portuguese language

Advances in automated essay grading over the last sixty years enabled its application in real scenarios, such as classrooms and high-stakes testing. The recognition of off-topic essays is one of the tasks addressed in automated essay grading. An essay is regarded as off-topic when the student does not develop the expected prompt-related concepts, sometimes purposely. Off-topic essays may receive a zero score in high-stake tests. An off-topic essay detection mechanism may be used in parallel or embedded in an automated essay grading system to improve its performance. In this context, the main goal of this study is to evaluate the existing approaches for automated off-topic essay detection. A previous systematic review of the literature showed some deficiencies in the state of the art, including: the low accuracy of current approaches, the use of artificial validation sets, and the lack of studies focused on the Portuguese language. In this study, the approaches found in the literature, originally proposed for the English language, were adapted for the Portuguese language and compared in an experiment using a public corpus of 2164 essays related to 111 prompts. The experiment used a set of artificial off-topic examples and the best performing algorithm achieved higher accuracy than that found in the literature for the English language (96.76% vs. 94.75%). The results presented suggest the application of off-topic essay detection mechanisms in the Brazilian educational context in order to benefit the student, with computer generated feedback, and educational institutions, regarding automated essay grading. Some suggestions for future research are presented, including the need to address the task of off-topic essay detection as a multiclass problem, and to reproduce the experiment with a larger and more representative set of real off-topic essay examples.

Categorized Image Classification using CNN Features with ECOC Framework

Image Classification technique is used to classify images into categories. In this study, an application is presented to examine category based image classification by combining Support Vector Machine with error correcting output codes (ECOC) framework. The ResNet50 used as Network architecture, our image dataset include caltech101 images from 9 categories (classes) which builds our classification task a multiclass problem. ECOC is a commonly used framework to model multiclass classification problem. We present one-verses-all coding design of ECOC and apply to SVM classifier. A pre-trained CNN (convolution neural network) is used for extracting image feature and as a classifier Multiclass Support Vector Machine is used. The extracted features are then passed for classification via ECOC approach. The final classification result predicts the class labels. The application is implemented in Matlab using pre-trained CNN. The prediction accuracy of each category is evaluated and presented. The experimental result shows an accuracy of 97.6%. Further experiments are carried out on different dataset which showed that best accuracy is achieved using CNN with ECOC for multiclass problem.

Comparing Multiclass, Binary, and Hierarchical Machine Learning Classification schemes for variable stars

ABSTRACT Upcoming synoptic surveys are set to generate an unprecedented amount of data. This requires an automatic framework that can quickly and efficiently provide classification labels for several new object classification challenges. Using data describing 11 types of variable stars from the Catalina Real-Time Transient Survey (CRTS), we illustrate how to capture the most important information from computed features and describe detailed methods of how to robustly use information theory for feature selection and evaluation. We apply three machine learning algorithms and demonstrate how to optimize these classifiers via cross-validation techniques. For the CRTS data set, we find that the random forest classifier performs best in terms of balanced accuracy and geometric means. We demonstrate substantially improved classification results by converting the multiclass problem into a binary classification task, achieving a balanced-accuracy rate of ∼99 per cent for the classification of δ Scuti and anomalous Cepheids. Additionally, we describe how classification performance can be improved via converting a ‘flat multiclass’ problem into a hierarchical taxonomy. We develop a new hierarchical structure and propose a new set of classification features, enabling the accurate identification of subtypes of Cepheids, RR Lyrae, and eclipsing binary stars in CRTS data.

Deep Generative Model with Supervised Latent Space for Text Classification

Recent advances in applying deep neural networks to Bayesian Modelling, sparked resurgence of in- terest in Variational Methods. Notably, the main contribution in this area is Reparametrization Trick introduced in [1] and [2]. VAE model [1], is unsupervised and therefore its application to classification is not optimal. In this work, we research the possibility to extend the model to supervised case. We first start with the model known as Supervised Variational Autoencoder that is researched in the literature in various forms [3] and [4]. We then modify objective function in such a way, that latent space can be better fitted to multiclass problem. Finally, we introduce a new method that uses information about classes to modify latent space, so it even better reflects differences between classes. All of this, will use only two dimensions. We will show, that mainstream classifiers applied to such a space, achieve significantly better performance than if applied to original datasets and VAE generated data. We also show, how our novel approach can be used to calculate better classification score, and how it can be used to generate data for a given class.

A HIERARCHICAL CLASSIFIER FOR MULTICLASS PROSTATE HISTOPATHOLOGY IMAGE GLEASON GRADING

Automated classification of prostate histopathology images includes the identification of multiple classes, such as benign and cancerous (grades 3 & 4). To address the multiclass classification problem in prostate histopathology images, breakdown approaches are utilized, such as one-versus-one (OVO) and one- versus-all (Ovall). In these approaches, the multiclass problem is decomposed into numerous binary subtasks, which are separately addressed. However, OVALL introduces an artificial class imbalance, which degrades the classification performance, while in the case of OVO, the correlation between different classes not regarded as a multiclass problem is broken into multiple independent binary problems. This paper proposes a new multiclass approach called multi-level (hierarchical) learning architecture (MLA). It addresses the binary classification tasks within the framework of a hierarchical strategy. It does so by accounting for the interaction between several classes and the domain knowledge. The proposed approach relies on the ‘divide-and-conquer’ principle by allocating each binary task into two separate subtasks; strong and weak, based on the power of the samples in each binary task. Conversely, the strong samples include more information about the considered task, which motivates the production of the final prediction. Experimental results on prostate histopathological images illustrated that the MLA significantly outperforms the Ovall and OVO approaches when applied to the ensemble framework. The results also confirmed the high efficiency of the ensemble framework with the MLA scheme in dealing with the multiclass classification problem.

Improving Predictions of Multiple Binary Models in ILP

Despite the success of ILP systems in learning first-order rules from small number of examples and complexly structured data in various domains, they struggle in dealing with multiclass problems. In most cases they boil down a multiclass problem into multiple black-box binary problems following the one-versus-one or one-versus-rest binarisation techniques and learn a theory for each one. When evaluating the learned theories of multiple class problems in one-versus-rest paradigm particularly, there is a bias caused by the default rule toward the negative classes leading to an unrealistic high performance beside the lack of prediction integrity between the theories. Here we discuss the problem of using one-versus-rest binarisation technique when it comes to evaluating multiclass data and propose several methods to remedy this problem. We also illustrate the methods and highlight their link to binary tree and Formal Concept Analysis (FCA). Our methods allow learning of a simple, consistent, and reliable multiclass theory by combining the rules of the multiple one-versus-rest theories into one rule list or rule set theory. Empirical evaluation over a number of data sets shows that our proposed methods produce coherent and accurate rule models from the rules learned by the ILP system of Aleph.