본 논문에서는 인공 신경망의 일종인 Extreme Learning Machine의 학습 알고리즘을 기반으로 하여 노이즈에 강한 특성을 보이는 퍼지 집합 이론을 이용한 새로운 패턴 분류기를 제안 한다. 기존 인공 신경망에 비해 학습속도가 매우 빠르며, 모델의 일반화 성능이 우수하다고 알려진 Extreme Learning Machine의 학습 알고리즘을 퍼지 패턴 분류기에 적용하여 퍼지 패턴 분류기의 학습 속도와 패턴 분류 일반화 성능을 개선 한다. 제안된 퍼지 패턴 분류기의 학습 속도와 일반화 성능을 평가하기 위하여, 다양한 머신 러닝 데이터 집합을 사용한다.

Extreme Learning Machine 기반 퍼지 패턴 분류기 설계

Integration of photovoltaics into power grids is difficult as solar energy is highly dependent on climate and geography; often fluctuating erratically. This causes penetrations and voltage surges, system instability, inefficient utilities planning and financial loss. Forecast models can help; however, time stamp, forecast horizon, input correlation analysis, data pre and post-processing, weather classification, network optimization, uncertainty quantification and performance evaluations need consideration. Thus, contemporary forecasting techniques are reviewed and evaluated. Input correlational analyses reveal that solar irradiance is most correlated with Photovoltaic output, and so, weather classification and cloud motion study are crucial. Moreover, the best data cleansing processes: normalization and wavelet transforms, and augmentation using generative adversarial network are recommended for network training and forecasting. Furthermore, optimization of inputs and network parameters, using genetic algorithm and particle swarm optimization, is emphasized. Next, established performance evaluation metrics MAE, RMSE and MAPE are discussed, with suggestions for including economic utility metrics. Subsequently, modelling approaches are critiqued, objectively compared and categorized into physical, statistical, artificial intelligence, ensemble and hybrid approaches. It is determined that ensembles of artificial neural networks are best for forecasting short term photovoltaic power forecast and online sequential extreme learning machine superb for adaptive networks; while Bootstrap technique optimum for estimating uncertainty. Additionally, convolutional neural network is found to excel in eliciting a model's deep underlying non-linear input-output relationships. The conclusions drawn impart fresh insights in photovoltaic power forecast initiatives, especially in the use of hybrid artificial neural networks and evolutionary algorithms.

A review and evaluation of the state-of-the-art in PV solar power forecasting:Techniques and optimization

Presently, educational institutions compile and store huge volumes of data, such as student enrolment and attendance records, as well as their examination results. Mining such data yields stimulating information that serves its handlers well. Rapid growth in educational data points to the fact that distilling massive amounts of data requires a more sophisticated set of algorithms. This issue led to the emergence of the field of educational data mining (EDM). Traditional data mining algorithms cannot be directly applied to educational problems, as they may have a specific objective and function. This implies that a preprocessing algorithm has to be enforced first and only then some specific data mining methods can be applied to the problems. One such preprocessing algorithm in EDM is clustering. Many studies on EDM have focused on the application of various data mining algorithms to educational attributes. Therefore, this paper provides over three decades long (1983–2016) systematic literature review on clustering algorithm and its applicability and usability in the context of EDM. Future insights are outlined based on the literature reviewed, and avenues for further research are identified.

A Systematic Review on Educational Data Mining

Abstract Class imbalance is often a problem in various real-world data sets, where one class (i.e. the minority class) contains a small number of data points and the other (i.e. the majority class) contains a large number of data points. It is notably difficult to develop an effective model using current data mining and machine learning algorithms without considering data preprocessing to balance the imbalanced data sets. Random undersampling and oversampling have been used in numerous studies to ensure that the different classes contain the same number of data points. A classifier ensemble (i.e. a structure containing several classifiers) can be trained on several different balanced data sets for later classification purposes. In this paper, we introduce two undersampling strategies in which a clustering technique is used during the data preprocessing step. Specifically, the number of clusters in the majority class is set to be equal to the number of data points in the minority class. The first strategy uses the cluster centers to represent the majority class, whereas the second strategy uses the nearest neighbors of the cluster centers. A further study was conducted to examine the effect on performance of the addition or deletion of 5 to 10 cluster centers in the majority class. The experimental results obtained using 44 small-scale and 2 large-scale data sets revealed that the clustering-based undersampling approach with the second strategy outperformed five state-of-the-art approaches. Specifically, this approach combined with a single multilayer perceptron classifier and C4.5 decision tree classifier ensembles delivered optimal performance over both small- and large-scale data sets.

分群式取樣法於類別不平衡問題之研究;Clustering-Based Under-sampling in Class Imbalanced Data

Due to sharp increases in data dimensions, working on every data mining or machine learning (ML) task requires more efficient techniques to get the desired results. Therefore, in recent years, researchers have proposed and developed many methods and techniques to reduce the high dimensions of data and to attain the required accuracy. To ameliorate the accuracy of learning features as well as to decrease the training time dimensionality reduction is used as a pre-processing step, which can eliminate irrelevant data, noise, and redundant features. Dimensionality reduction (DR) has been performed based on two main methods, which are feature selection (FS) and feature extraction (FE). FS is considered an important method because data is generated continuously at an ever-increasing rate; some serious dimensionality problems can be reduced with this method, such as decreasing redundancy effectively, eliminating irrelevant data, and ameliorating result comprehensibility. Moreover, FE transacts with the problem of finding the most distinctive, informative, and decreased set of features to ameliorate the efficiency of both the processing and storage of data. This paper offers a comprehensive approach to FS and FE in the scope of DR. Moreover, the details of each paper, such as used algorithms/approaches, datasets, classifiers, and achieved results are comprehensively analyzed and summarized. Besides, a systematic discussion of all of the reviewed methods to highlight authors' trends, determining the method(s) has been done, which significantly reduced computational time, and selecting the most accurate classifiers. As a result, the different types of both methods have been discussed and analyzed the findings.

/pdf/a-comprehensive-review-of-dimensionality-reduction-4vd9o1ltn6.pdf

A Comprehensive Review of Dimensionality Reduction Techniques for Feature Selection and Feature Extraction

While training a model with data from a dataset, we have to think of an ideal way to do so. The training should be done in such a way that while the model has enough instances to train on, they should not over-fit the model and at the same time, it must be considered that if there are not enough instances to train on, the model would not be trained properly and would give poor results when used for testing. Accuracy is important when it comes to classification and one must always strive to achieve the highest accuracy, provided there is not trade off with inexcusable time. While working on small datasets, the ideal choices are k-fold cross-validation with large value of k (but smaller than number of instances) or leave-one-out cross-validation whereas while working on colossal datasets, the first thought is to use holdout validation, in general. This article studies the differences between the two validation schemes, analyzes the possibility of using k-fold cross-validation over hold-out validation even on large datasets. Experimentation was performed on four large datasets and results show that till a certain threshold, k-fold cross-validation with varying value of k with respect to number of instances can indeed be used over hold-out validation for quality classification.

Analysis of k-Fold Cross-Validation over Hold-Out Validation on Colossal Datasets for Quality Classification

Data preprocessing is an important step for designing classification model. Normalization is one of the preprocessing techniques used to handle the out-of-bounds attributes. This work develops 14 classification models using different learning algorithms for dynamic selection of normalization technique. This work extracts 12 data complexity measures for 48 datasets drawn from the KEEL dataset repository. Each of these datasets is normalized using min–max and z-score normalization technique. G-mean index is estimated for these normalized datasets using Gaussian Kernel Extreme Learning Machine (KELM) in order to determine the best-suited normalization technique. The data complexity measures along with the best-suited normalization technique are used as an input for developing the aforementioned dynamic models. These models predict the best suitable normalization technique based on the estimated data complexity measures of the dataset. The result shows that the model developed using Gaussian Kernel ELM (KELM) and Support Vector Machine (SVM) give promising results for most of the evaluated classification problems.

Dynamic selection of normalization techniques using data complexity measures

Class imbalance learning using UnderBagging based kernelized extreme learning machine

Imbalanced learning is one of the substantial challenging problems in the field of data mining. The datasets that have skewed class distribution pose hindrance to conventional learning methods. Conventional learning methods give the same importance to all the samples. This leads to biased accuracy, which favors the majority classes. Several classifiers have been designed to tackle the class imbalance problems. Weighted kernel-based SMOTE (WKSMOTE) is a recently proposed method, which employs the minority oversampling in kernel space to tackle the class imbalance problem. Motivated by WKSMOTE, this work proposes a novel SMOTE based class-specific extreme learning machine (SMOTE-CSELM), a variant of class-specific extreme learning machine (CS-ELM), which exploits the benefit of both the minority oversampling and the class-specific regularization. For minority oversampling, this work uses synthetic minority oversampling technique (SMOTE). It increases the significance of the minority class samples for determining the decision region of the classifiers. The proposed method has comparable computational complexity than the weighted extreme learning machine (WELM) for imbalanced learning. The extensive experimental results evaluated on the real-world benchmark datasets demonstrate the efficacy of our proposed method.

Sanyam Shukla

Papers

Analysis of k-Fold Cross-Validation over Hold-Out Validation on Colossal Datasets for Quality Classification

Dynamic selection of normalization techniques using data complexity measures

Class imbalance learning using UnderBagging based kernelized extreme learning machine

SMOTE based class-specific extreme learning machine for imbalanced learning

Pre-trained convolutional neural networks as feature extractors for diagnosis of breast cancer using histopathology