Gaussian ERP Kernel Classifier for Pulse Waveforms Classification

On The Marriage of Lp-norms and Edit Distance

Many researchers have been attracted towards searching of similar moving objects' trajectories due to their myriad applications. Similarity of trajectories is determined mainly by closeness or shape parameters. Closeness is measured by distance measures like Dynamic Time Warping (DTW), Edit Distance with Real Penalty (ERP), Edit Distance with Real Sequence (EDR) and Longest Common Subsequence (LCSS). These edit distance measures support scaling and translation property, but do not support rotation invariant property. Rotation Invariant (RI) distance measure supports all three properties and hence it is considered to be superior compared to other edit distance measures. However, it has two main drawbacks. The first one is, RI does not compares trajectories based on the shape. The second drawback is, RI is not robust to noise and hence produces poor results. In order to eliminate drawbacks faced by existing measures, in this paper, we have proposed Polygon Based Distance (PBD) measure to compare trajectories for similarity. Our proposed PBD distance measure supports scaling and translation properties like other measures. The main advantages of PBD distance measure are, it is invariant of rotation, it compares the trajectories based on the shape and it is robust to noise. We have performed experimental study on real time and synthetic datasets. The average accuracy of DTW is 21.67, RI is 35 and PBD is 58.33. Since PBD measure compares the trajectories based on shape and is robust to noise, its accuracy is higher compared to DTW and RI distance measures.

In traditional Chinese pulse diagnosis (TCPD), diseases of internal organs can be detected by recognizing pulse waveform patterns of wrist radial arterial. However pulse waveform analysis, for which Doppler diagnosis is a powerful tool, is limited to cardiovascular diseases. This paper tries to fill the gap between TCPD and Doppler diagnosis by applying signal analysis and pattern recognition technologies to Doppler blood flow signals (DBFS's) of wrist radial arterial, which are recorded from both hands of healthy people, gastritis and cholecystitis patients. DBFS's are classified using the features proposed by an L2-soft margin support vector machine (L2-SVM): five clinical Doppler parameters (DP), wavelet energies (WE), wavelet packet energies (WPE), and piecewise axially integrated bispectra (PAIB). 5-fold cross validation is used for performance evaluation. The sick are differentiated from the healthy with an accuracy of about 80% using DP, WE and WPE, while the classification rate between gastritis and cholecystitis reaches 100%. Using PAIB, ether two groups of subjects are classified with accuracy greater than 93%. Gastritis is more accurately recognized than cholecystitis, while the latter is recognized with a higher accuracy on data from the left hand than right. Though the sample size is relatively small, we still argue that the methods proposed here are effective and could serve as an assisstive tool for TCPD.

This paper reviews recent achievements in quantitative analyses of modern research on Traditional Chinese Pulse Diagnosis (TCPD). In order to demystify TCPD and prove its efficiency, some fundamental knowledge such as concepts, diagnosis methods, and standard pulse patterns are discussed. We review the modern research on TCPD mainly from 4 aspects: objectification of TCPD, analyses of pulse waveform, research into the mechanism of pulse formation, clinic observations, and comparisons on pulse images. For each of these aspects, general background information and a brief explanation on them are given. It is especially important to distinguish the pulse images based on Traditional Chinese Medicine (TCM) and the sphygmogram based on Western medicine. Our pulse acquisition system is presented as an example. Furthermore, typical pulse waveforms in our pulse database and their results processed by modern signal processing methods such as cepstrum, short-time Fourier transforms (STFT), and wavelet transf...

Background: Data on early childhood disease collected in clinics has accumulated into big data. Those data can be used for classification of early childhood diseases to help medical staff in diagnosing diseases that attack early childhoods.Objective: This study aims to apply Principal Component Analysis (PCA) and K-Nearest Neighbor (K-NN) Classifier for the classification of early childhood diseases.Methods: Data analysis was performed using PCA to obtain variables that had a major influence on the classification of early childhood diseases. PCA was done by observing the correlation between variables and eliminating variables that have little influence on classification. Furthermore, data on early childhood disease was classified using the K-Nearest Neighbor Classifier method.Results: The results of system evaluation using 150 test data indicated that the classification system by applying PCA and KNN Classifier had an accuracy value of 86%.Conclusion: PCA can be used to reduce the number of variables involved so that it can improve system performance in terms of efficiency. In addition, the application of PCA and KNN can also improve accuracy in the classification of early childhood diseases.

Solid waste bin level detection using gray level co-occurrence matrix feature extraction approach

Spectrum sensing use current primary user (PU) received signal to detect PU channel state at current time, however PU channel state in near future using previous detected channel state information can secure an efficient utilization of unoccupied spectrum, also can help in preventing harmful interference with PU and solving problem of latency between spectrum sensing and data transmission which caused by hardware implementation. In this paper, an algorithm Based on K-Nearest Neighbor (KNN) Classifier and Kalman Filter is proposed to detect and predict future of primary user channel state the spectrum holes. The new algorithm based on training KNN classifier over a set containing some frame time slots energy test statistics along with their corresponding decisions about presence or absence of primary user transmission, so as to predict decisions for new unclassified time slot, then utilize Kalman filter to track PU primary user channel state by next expected spectrum holes based on previously detected hole. Experimental results show that proposed approach is efficient and easy to implement for detecting and tracking predicting PU channel state. KEYWORD: Channel state tracking; Channel state prediction; K-nearest neighbor classifier; Kalman Filter; Spectrum sensing time slot

Recently, computer and smartphone devices have been increasingly employed in various life aspects. Human Activity Recognition (HAR) is one of the most emerging fields that can be clearly noticed in terms of computer related researches. This study proposes a classification model to recognize human six basic activities based on smartphone embedded sensors. For feature selection, we utilized three optimization search algorithms which are: Particle Swarm Optimization (PSO), Greedy, and Genetic algorithm. K-nearest neighbors (KNN) classifier, with Manhattan distance function, is adopted to measure the models performance on two publicly available well-known datasets (i.e., WISDM and UCI-HAR). All experiments are achieved using 10-folds cross validation technique, and the results are evaluated and compared using different metrics, for instance, accuracy, root mean square error, f-measure, and number of features. The experimental results proved that PSO based KNN classifier presents the best performance compared to the other two algorithms, in addition to the baseline KNN classifier. Moreover, the proposed model outperforms those state-of-the-art models presented in related literature regarding the UCI-HAR dataset with performance rate of 98.89%, and close to the best performance achieved on WISDM dataset with performance rate of 91.07%.

Abstract- This study uses a K-Nearest Neighbours (KNN) classifier to detect teenage cardiac disease. Years of age, Sexual orientation, Chest Ache category, Ambient Heart pressure, cholesterol levels, Highest Cardiac Level, and Exercise-Induced Angina are clinically important and interpretable aspects of the cardiovascular disease dataset used to create the model. These qualities were chosen to improve model interpretability for doctors and laypeople. Data preparation encoded categorical variables and standardised features for model optimisation. Histograms, charts with bars, and scatter plots were used to explore feature distribution and heart disease status. The kernel nearest neighbour (KNN) model was trained with k=11 neighbours and tested using precision, a matrix of confusion, classifying report, ROC spectrum, and precision-recall curve, proving predictive power. To select the most influential predictors, feature importance was permutation-based. A Flask web application lets users input health parameters to forecast heart disease risk. The software generates personalised, patient-friendly health information using OpenAI's API and generates PDF reports. The technology keeps prognosis records in an information system for future reference. This approach uses data-driven modelling, interpretability, and user-friendliness to help adolescents recognise and understand heart disease risk. Keywords - Adolescent heart disease identification, risk factors for heart disease, KNN classifier, artificial intelligence, data preprocessing, feature relevance, model assessment. Flask web app, OpenAI integration into the API, health insights, PDF reports, early diagnosis, and healthcare information visualisation.

Underwater garbage affects the life of marine creatures and the entire ecosystem. Detecting underwater garbage is an important research area. In this study, a method is proposed to detect underwater garbage. The open-access Trash-ICRA19 dataset was used to implement the proposed method. The data set cropping process was applied and a data set consisting of 11060 images in total was obtained. These images were converted to 200×200 pixels using preprocessing. By applying the Directed Gradient Histogram (HOG) algorithm, 11060×900 feature vectors were obtained. The resulting feature vectors were then calculated using KNN (K Nearest Neighbor Algorithm), DT (Decision Tree), LD (Linear Discriminant), NB (Naive Bayes), and SVM (Support Vector Machine) classifiers. The results obtained showed that 97.78% accuracy was obtained when the KNN classifier was used in this method. The use of only feature extractors and classifiers in the proposed method shows that the method is lightweight. It has low computational complexity compared to existing studies in the literature. Moreover, according to its performance results, it is more successful than the methods in the literature.

This paper presents the combination of electroencephalogram (EEG) power spectrum ratio and Spectral Centroids techniques to extract unique features for human stress from EEG signals. The combination of these techniques was able to improve the k-NN (k-Nearest Neighbor) classifier accuracy to detect and classify human stress from two cognitive states, Close-eye (CE) and Open-eye (OE). The EEG power spectrum in term of Energy Spectral Density (ESD) for each frequency bands (Delta, Theta, Alpha and Beta) was calculated. The ratio of EEG power spectrum and the average value of Spectral Centroids were selected as features to k-Nearest Neighbor (k-NN). The training and testing of the classifier were evaluated at 50:50 ratios and 70:30 ratios. The results showed that the combination of EEG power spectrum and Spectral Centroids techniques with the training and testing of k-NN set at 70:30 able to detect and classify the unique features for human stress at 88.89% accuracy.

Traditional Chinese Pulse Diagnosis (TCPD), one of the four diagnostic methods of Traditional Chinese Medicine (TCM), had been proved to be clinically valid in Chinese Medicine history. Different from most previous work which focused on the diagnosis of cardiovascular diseases, this paper further investigated the possibility of diagnosing cholecystitis and nephrotic syndrome using the pulse waveform data. After the pre-processing, the pulse waveform signals were decomposed into a given level by Wavelet Packet Transform and the best basis was picked out by Shannon entropy criterion. Then, subband energies contained in the best basis were extracted as features and the support vector machine classifiers were trained. Experimental results indicated that the proposed method can effectively discriminate these two kinds of diseases.

The K-Nearest Neighbour (KNN) classification is a standard method in pattern recognition and is especially marked out by its simplicity [128, 148, 305, 370, 374, 379, 389].

This paper presents automatic detection and localization of myocardial infarction (MI) using K-nearest neighbor (KNN) classifier. Time domain features of each beat in the ECG signal such as T wave amplitude, Q wave and ST level deviation, which are indicative of MI, are extracted from 12 leads ECG. Detection of MI aims to classify normal subjects without myocardial infarction and subjects suffering from Myocardial Infarction. For further investigation, Localization of MI is done to specify the region of infarction of the heart. Total 20,160 ECG beats from PTB database available on Physio-bank is used to investigate the performance of extracted features with KNN classifier. In the case of MI detection, sensitivity and specificity of KNN is found to be 99.9% using half of the randomly selected beats as training set and rest of the beats for testing. Moreover, Arif-Fayyaz pruning algorithm is used to prune the data which will reduce the storage requirement and computational cost of search. After pruning, sensitivity and specificity are dropped to 97% and 99.6% respectively but training is reduced by 93%. Myocardial Infarction beats are divided into ten classes based on the location of the infarction along with one class of normal subjects. Sensitivity and Specificity of above 90% is achieved for all eleven classes with overall classification accuracy of 98.8%. Some of the ECG beats are misclassified but interestingly these are misclassified to those classes whose location of infarction is near to the true classes of the ECG beats. Pruning is done on the training set for eleven classes and training set is reduced by 70% and overall classification accuracy of 98.3% is achieved. The proposed method due to its simplicity and high accuracy over the PTB database can be very helpful in correct diagnosis of MI in a practical scenario.

Attentional bias in MDD: ERP components analysis and classification using a dot-probe task