Post Classifier Research Articles

Online Handwritten Telugu Character Recognition Using Normalized Differential Chain Code Feature

In Character Recognition, the Feature extraction has encompassed a well-known role. Here, Feature Extraction centered on Chain code (CC) is implemented. CC encodes every stroke with a string of numbers, in which every number signifies a specific direction wherein the subsequent point on the stroke is present. CC centered feature safeguard information and permits reasonable data to decrease. Disparate CC can signify the same shape since the CC is reliant on starting point. So here, Starting Point and rotation invariant feature extraction technique using Normalized Differential Chain Code (NDCC) is proposed. A two-stage classifier is employed for classification. Here, the NDCC feature is utilized in the pre-classifier and pre-processed (x,y) coordinates are used in the post classifier. In both stages K-NN classifier is used. This feature is verified in HP-Lab data that is present in the UNIPEN format. Investigational outcomes proved that the proposed feature enhances recognition accuracy over the selected dataset.

Signal-piloted processing and machine learning based efficient power quality disturbances recognition.

Significant losses can occur for various smart grid stake holders due to the Power Quality Disturbances (PQDs). Therefore, it is necessary to correctly recognize and timely mitigate the PQDs. In this context, an emerging trend is the development of machine learning assisted PQDs management. Based on the conventional processing theory, the existing PQDs identification is time-invariant. It can result in a huge amount of unnecessary information being collected, processed, and transmitted. Consequently, needless processing activities, power consumption and latency can occur. In this paper, a novel combination of signal-piloted acquisition, adaptive-rate segmentation and time-domain features extraction with machine learning tools is suggested. The signal-piloted acquisition and processing brings real-time compression. Therefore, a remarkable reduction can be secured in the data storage, processing and transmission requirement towards the post classifier. Additionally, a reduced computational cost and latency of classifier is promised. The classification is accomplished by using robust machine learning algorithms. A comparison is made among the k-Nearest Neighbor, Naïve Bayes, Artificial Neural Network and Support Vector Machine. Multiple metrics are used to test the success of classification. It permits to avoid any biasness of findings. The applicability of the suggested approach is studied for automated recognition of the power signal’s major voltage and transient disturbances. Results show that the system attains a 6.75-fold reduction in the collected information and the processing load and secures the 98.05% accuracy of classification.

Lightme: analysing language in internet support groups for mental health.

Assisting moderators to triage harmful posts in Internet Support Groups is relevant to ensure its safe use. Automated text classification methods analysing the language expressed in posts of online forums is a promising solution. Natural Language Processing and Machine Learning technologies were used to build a triage post classifier using a dataset from Reachout.com mental health forum for young people. When comparing with the state-of-the-art, a solution mainly based on features from lexical resources, received the best classification performance for the crisis posts (52%), which is the most severe class. Six salient linguistic characteristics were found when analysing the crisis post; (1) posts expressing hopelessness, (2) short posts expressing concise negative emotional responses, (3) long posts expressing variations of emotions, (4) posts expressing dissatisfaction with available health services, (5) posts utilising storytelling, and (6) posts expressing users seeking advice from peers during a crisis. It is possible to build a competitive triage classifier using features derived only from the textual content of the post. Further research needs to be done in order to translate our quantitative and qualitative findings into features, as it may improve overall performance.

ASSESSMENT OF EPILEPSY CLASSIFICATION USING TECHNIQUES SUCH AS SINGULAR VALUE DECOMPOSITION, APPROXIMATE ENTROPY, AND WEIGHTED K-NEAREST NEIGHBORS MEASURES

ABSTRACTObjective: The main aim of this research is to reduce the dimension of the epileptic Electroencephalography (EEG) signals and then classify it usingvarious post classifiers. For the evaluation and easy treatment of neurological diseases, EEG signals are used. The reflection of the electrical activitiesof the human brain is obtained by the measurement of potentials in EEG. To study and explore the brain functions in an exhaustive manner, EEG is usedby both physicians and scientists. The study of the electrical activity of the brain which is done through EEG recording is a vital tool for the diagnosis ofmany neurological diseases which include epilepsy, sleep disorders, injuries in head, dementia etc. One of the most commonly occurring and prevalentneurological disorders is epilepsy and it is easily characterized by recurrent seizures.Methods: This paper employs the concept of dimensionality reduction concepts like Fuzzy Mutual Information (FMI), Independent ComponentAnalysis (ICA), Linear Graph Embedding (LGE), Linear Discriminant Analysis (LDA) and finally Variational Bayesian Matrix Factorization (VBMF).The epilepsy risk levels are also classified using post classifiers like Singular Value Decomposition (SVD), Approximate Entropy (ApEn) and WeightedKNN (W-KNN) classifiers.Results: The highest accuracy is obtained when LDA is combined with Weighted KNN (W-KNN) Classifiers and it is of 97.18%. Conclusion: Thus the EEG signals not only represent the brain function but also the status of the whole body. The best result obtained was whenLDA is engaged as a dimensionality reduction technique followed by the usage of the W-KNN as post classifier for the classification of epilepsy risklevels from EEG signals. Future work may incorporate the possible usage of different dimensionality reduction techniques with various other types ofclassifiers for the perfect classification of epilepsy risk levels from EEG signals.Keywords: FMI, ICA, LGE, LDA, W-KNN, EEG

Variational Bayesian Matrix Factorization and Certain Post Classifiers for Classification of Epilepsy from EEG Signals

The main aim of this paper is to employ Variational Bayesian Matrix Factorization (VBMF) as a dimensionality reduction technique followed by the Gaussian Mixture Model (GMM), Genetic Algorithm (GA) and Naive Bayes Classifier (NBC) as post classifiers for the classification of epilepsy risk levels from Electroencephalography (EEG) Signals. Since epilepsy is one of the serious disorders of the brain which is characterized by frequent and recurrent seizures, the detection and classification of it seems to be very important. Using the EEG signals, the epileptic seizures can be analyzed because it aids in the recording, diagnosing and for treating other neurological disorders. In this paper, the results are analyzed and compared in terms of sensitivity, specificity, time delay, quality values, performance index and accuracy.

A real time experimental setup for classification of epilepsy risk levels

The objective of this paper is to analyze the performance of singular value decomposition, expectation maximization, and Elman Neural Networks in optimization of code converter outputs in the classification of epilepsy risk levels from EEG (electroencephalogram) signals. The signal parameters such as the total number of positive and negative peaks, spikes and sharp waves, their duration etc., were extracted using morphological operators and wavelet transforms. Code converters were considered as a level one classifier. Code converters were found to have a performance index and quality value of 33.26 and 12.74, respectively, which is low. Consequently, for the EEG signals of 20 patients, the post classifiers were applied across 3 epochs of 16 channels. After having made a comparative study of different architectures, SVD was found to be the best post classifier as it marked a performance index of 89.48 and a quality value of 20.62. Elman neural network also exhibits good performance metrics than SVD in the morphological operator based feature extraction method.

A Novel Approach of Eye Detection Based on Haar-Like Feature and SVM

Recognizing the importance of eye detection technology in face detection and recognition, facial expression recognition, driver fatigue detection, and so on, the current paper presents a novel approach to eye detection in images based on AdaBoost and the Support Vector Machine (SVM). The proposed method can accurately detect eyes with different postures, facial expressions, skin colors, and glasses in a simple background image. First, a Haar-like feature AdaBoost classifier based on the developed training samples is used to detect eyes in images. An SVM post classifier trained by coordinates of the true and pseudo eye rectangles is then used to optimize the results. Experimental results show that the proposed method is fast, robust, and has high generalization capability.

Performance Analysis of Patient Specific Elman-Chaotic Optimization Model for Fuzzy Based Epilepsy Risk Level Classification from Eeg Signals

Abstract This paper aims to analyze the optimization of Epilepsy risk levels from EEG signals using Fuzzy based Elman-Chaotic Optimization. The EEG (Electroencephalogram) signals of twenty patients are collected from Sri Ramakrishna Hospitals at Coimbatore. The raw EEG signals are sampled and various parameters like energy, energy, variance, peaks, sharp and spike waves, duration, events and covariance. The fuzzy techniques are applied as a first level classifier to classify the risk levels of epilepsy by converting the EEG signal parameters in to code patterns by fuzzy systems. Elman-Chaotic optimization is identified as post classifiers on the classified data to obtain the optimized risk level that characterizes the patient’s epilepsy risk level. This classification provides a better way of treating the epileptic patients. This project aims to safeguard a patient’s life when critical situation occurs. Future scope is to design an embedded system which collects the raw EEG signals from the brain and directly gives the level of epilepsy. It will make the neural surgeons to give appropriate remedial measures.

Speech recognition system having a quantizer using a single robust codebook designed to multiple signal to noise ratios

In one embodiment, a speech recognition system is organized with a fuzzy matrix quantizer with a single codebook representing u codewords. The single codebook is designed with entries from u codebooks which are designed with respective words at multiple signal to noise ratio levels. Such entries are, in one embodiment, centroids of clustered training data. The training data is, in one embodiment, derived from line spectral frequency pairs representing respective speech input signals at various signal to noise ratios. The single codebook trained in this manner provides a codebook for a robust front end speech processor, such as the fuzzy matrix quantizer, for training a speech classifier such as a u hidden Markov models and a speech post classifier such as a neural network. In one embodiment, a fuzzy Viterbi algorithm is used with the hidden Markov models to describe the speech input signal probabilistically.