Machine Learning based Identification and Classification of Field-Operation caused Solar Panel Failures observed in Electroluminescence Images

Failure or degradation effects lead to power losses in solar panels during their field operation and are identified commonly by electroluminescence imaging. Failures like potential induced degradation and light and enhanced temperature induced degradation require an identification of the electroluminescence pattern over the entire solar panel. As the manual process of analysing patterns is prone to error, we seek for an automatic detection of these failure types. We predict automatically the failure types potential induced degradation and light and enhanced temperature induced degradation by adopting the principle component analysis method in combination with a k-nearest neighbour classifier.<br>

Predicting Blood Glucose Level in Malaysian Women Based on Glutathione and Anthropometric Parameters

Blood glucose is conventionally determined by the level of sugar present in our blood. Lesser known to the public that antioxidants in our body are also said to influence the level of blood glucose. Glutathione (GSH) as the main antioxidant parameter in our body helps in reducing the production of oxidative stress caused by a high blood glucose level. Particularly in women, high antioxidant activities are reported due to the presence of oestrogen hormone. However, in Malaysia limited study was done on the significance of GSH in influencing the blood glucose level. Thus, this study focuses on finding the significance of GSH and some other health predictors in affecting the blood glucose level of women volunteers. This study was carried out on 118 Malaysian women volunteers and blood samples were collected for GSH analysis and blood glucose. All data were trained and tested for the development of prediction models in classifying the blood glucose into normal and abnormal levels. The model construction is using three different classifiers: namely logistic regression, k-nearest neighbour classifier and decision tree. Five predictors that were used are GSH, weight, body mass index (BMI), waist-hip ratio (WHR) and groups (oral supplementation dosage). Results showed all predictors are significantly correlated with the blood glucose level at p < 0.10. The model with a combination of GSH, BMI, WHR, weight and supplementation dosage (groups) as predictors gave the best performance. The k-nearest neighbour classifier model displays the best accuracy (84.7%) in predicting the normal and abnormal level of blood glucose. This finding shows that by altering the amount of GSH via oral supplementation and other significant predictors in women, there are chances to modify the blood glucose level from abnormal to normal

Machine Learning based Identification and Classification of Field-Operation caused Solar Panel Failures observed in Electroluminescence Images

Failure or degradation effects lead to power losses in solar panels during their field operation and are identified commonly by electroluminescence imaging. Failures like potential induced degradation and light and enhanced temperature induced degradation require an identification of the electroluminescence pattern over the entire solar panel. As the manual process of analysing patterns is prone to error, we seek for an automatic detection of these failure types. We predict automatically the failure types potential induced degradation and light and enhanced temperature induced degradation by adopting the principle component analysis method in combination with a k-nearest neighbour classifier.<br>

Use of the K-Nearest Neighbour Classifier in Wear Condition Classification of a Positive Displacement Pump

This paper presents a learning system with a K-nearest neighbour classifier to classify the wear condition of a multi-piston positive displacement pump. The first part reviews current built diagnostic methods and describes typical failures of multi-piston positive displacement pumps and their causes. Next is a description of a diagnostic experiment conducted to acquire a matrix of vibration signals from selected locations in the pump body. The measured signals were subjected to time-frequency analysis. The signal features calculated in the time and frequency domain were grouped in a table according to the wear condition of the pump. The next step was to create classification models of a pump wear condition and assess their accuracy. The selected model, which best met the set criteria for accuracy assessment, was verified with new measurement data. The article ends with a summary.

Human Authentication using Face, Voice and Fingerprint Biometrics

Multimodal biometric approaches are growing in importance for personal verification and identification, since they provide better recognition results and hence improve security compared to biometrics based on a single modality. In this project, we present a multimodal biometric system that is based on the fusion of face, voice and fingerprint biometrics. For face recognition, we employ Haar Cascade Algorithm, while minutiae extraction is used for fingerprint recognition and we will be having a stored code word for the voice authentication, if any of these two authentication becomes true, the system consider the person as authorized person. Fusion at matching score level is then applied to enhance recognition performance. In particular, we employ the product rule in our investigation. The final identification is then performed using a nearest neighbour classifier which is fast and effective. Experimental results confirm that our approach achieves excellent recognition performance, and that the fusion approach outperforms biometric identification based on single modalities.

k-Nearest Neighbour Classifiers - A Tutorial

Perhaps the most straightforward classifier in the arsenal or Machine Learning techniques is the Nearest Neighbour Classifier—classification is achieved by identifying the nearest neighbours to a query example and using those neighbours to determine the class of the query. This approach to classification is of particular importance, because issues of poor runtime performance is not such a problem these days with the computational power that is available. This article presents an overview of techniques for Nearest Neighbour classification focusing on: mechanisms for assessing similarity (distance), computational issues in identifying nearest neighbours, and mechanisms for reducing the dimension of the data. This article is the second edition of a paper previously published as a technical report [16]. Sections on similarity measures for time-series, retrieval speedup, and intrinsic dimensionality have been added. An Appendix is included, providing access to Python code for the key methods.

A novel binary k-mer approach for classification of coding and non-coding RNAs across diverse species

Classification among coding (CDS) and non-coding RNA (ncRNA) sequences is a challenge and several machine learning models have been developed for the same. Since the frequency of curated coding sequences is many-folds as compared to that of the ncRNAs, we devised a novel approach to work with the complete datasets from fifteen diverse species. In our proposed novel binary approach, we replaced all the A,T with 0 and G,C with 1 to obtain a binary form of coding and ncRNAs. The k-mer analysis of these binary sequences revealed that the frequency of binary patterns among the coding and ncRNAs can be used as features to distinguish among them. Using insights from these distinguishing frequencies, we used k-nearest neighbour classifier to classify among them. Our strategy is not only time-efficient but leads to significantly increased performance metrics including Matthews correlation coefficient (MCC) for some species like P. paniscus, M. mulatta, M. lucifugus, G. gallus, C. japonica, C. abingdonii, A. carolinensis, D. melanogaster and C. elegans when compared with the conventional ATGC approach. Additionally, we also show that the values of MCC obtained for diverse species tested on the model based on H. sapiens correlated with the geological evolutionary timeline thereby further strengthening our approach. Therefore, we propose that CDS and ncRNAs can be efficiently classified using 2-character frequency as compared to 4-character frequency of ATGC approach. Thus, our highly efficient binary approach can replace the more complex ATGC approach successfully.

Learning-free pattern detection for manuscript research:

Automatic pattern detection has become increasingly important for scholars in the humanities as the number of manuscripts that have been digitised has grown. Most of the state-of-the-art methods used for pattern detection depend on the availability of a large number of training samples, which are typically not available in the humanities as they involve tedious manual annotation by researchers (e.g. marking the location and size of words, drawings, seals and so on). This makes the applicability of such methods very limited within the field of manuscript research. We propose a learning-free approach based on a state-of-the-art Naïve Bayes Nearest-Neighbour classifier for the task of pattern detection in manuscript images. The method has already been successfully applied to an actual research question from South Asian studies about palm-leaf manuscripts. Furthermore, state-of-the-art results have been achieved on two extremely challenging datasets, namely the AMADI_LontarSet dataset of handwriting on palm leaves for word-spotting and the DocExplore dataset of medieval manuscripts for pattern detection. A performance analysis is provided as well in order to facilitate later comparisons by other researchers. Finally, an easy-to-use implementation of the proposed method is developed as a software tool and made freely available.

IoT-cloud based healthcare model for COVID-19 detection: an enhanced k-Nearest Neighbour classifier based approach

COVID - 19 affected severely worldwide. The pandemic has caused many causalities in a very short span. The IoT-cloud-based healthcare model requirement is utmost in this situation to provide a better decision in the covid-19 pandemic. In this paper, an attempt has been made to perform predictive analytics regarding the disease using a machine learning classifier. This research proposed an enhanced KNN (k NearestNeighbor) algorithm eKNN, which did not randomly choose the value of k. However, it used a mathematical function of the dataset’s sample size while determining the k value. The enhanced KNN algorithm eKNN has experimented on 7 benchmark COVID-19 datasets of different size, which has been gathered from standard data cloud of different countries (Brazil, Mexico, etc.). It appeared that the enhanced KNN classifier performs significantly better than ordinary KNN. The second research question augmented the enhanced KNN algorithm with feature selection using ACO (Ant Colony Optimization). Results indicated that the enhanced KNN classifier along with the feature selection mechanism performed way better than enhanced KNN without feature selection. This paper involves proposing an improved KNN attempting to find an optimal value of k and studying IoT-cloud-based COVID - 19 detection.