Crack Fault Classification for Planetary Gearbox Based on Feature Selection Technique and K-means Clustering Method

During the condition monitoring of a planetary gearbox, features are extracted from raw data for a fault diagnosis.However, different features have different sensitivity for identifying different fault types, and thus, the selection of a sensitive feature subset from an entire feature set and retaining as much of the class discriminatory information as possible has a directly effect on the accuracy of the classification results. In this paper, an improved hybrid feature selection technique(IHFST) that combines a distance evaluation technique(DET), Pearson’s correlation analysis, and an ad hoc technique is proposed. In IHFST, a temporary feature subset without irrelevant features is first selected according to the distance evaluation criterion of DET, and the Pearson’s correlation analysis and ad hoc technique are then employed to find and remove redundant features in the temporary feature subset, respectively, and hence,a sensitive feature subset without irrelevant or redundant features is selected from the entire feature set. Further, the k-means clustering method is applied to classify the different kinds of health conditions. The effectiveness of the proposed method was validated through several experiments carried out on a planetary gearbox with incipient cracks seeded in the tooth root of the sun gear, planet gear, and ring gear. The results show that the proposed method can successfully distinguish the different health conditions of a planetary gearbox, and achieves a better classification performance than other methods. This study proposes a sensitive feature subset selection method that achieves an obvious improvement in terms of the accuracy of the fault classification.

A Hybrid K-Means-GRA-SVR Model Based on Feature Selection for Day-Ahead Prediction of Photovoltaic Power Generation

In order to ensure that the large-scale application of photovoltaic power generation does not affect the stability of the grid, accurate photovoltaic (PV) power generation forecast is essential. A short-term PV power generation forecast method using the combination of K-means++, grey relational analysis (GRA) and support vector regression (SVR) based on feature selection (Hybrid Kmeans-GRA-SVR, HKGSVR) was proposed. The historical power data were clustered through the multi-index K-means++ algorithm and divided into ideal and non-ideal weather. The GRA algorithm was used to match the similar day and the nearest neighbor similar day of the prediction day. And selected appropriate input features for different weather types to train the SVR model. Under ideal weather, the average values of MAE, RMSE and R2 were 0.8101, 0.9608 kW and 99.66%, respectively. And this method reduced the average training time by 77.27% compared with the standard SVR model. Under non-ideal weather conditions, the average values of MAE, RMSE and R2 were 1.8337, 2.1379 kW and 98.47%, respectively. And this method reduced the average training time of the standard SVR model by 98.07%. The experimental results show that the prediction accuracy of the proposed model is significantly improved compared to the other five models, which verify the effectiveness of the method.

Feature Selection for Image Classification Based on a New Ranking Criterion

In this paper, a feature selection method combining the reliefF and SVM-RFE algorithm is proposed. This algorithm integrates the weight vector from the reliefF into SVM-RFE method. In this method, the reliefF filters out many noisy features in the first stage. Then the new ranking criterion based on SVM-RFE method is applied to obtain the final feature subset. The SVM classifier is used to evaluate the final image classification accuracy. Experimental results show that our proposed relief- SVM-RFE algorithm can achieve significant improvements for feature selection in image classification.

Genetic Algorithm Combined with the K-Means Algorithm:A Hybrid Technique for Unsupervised Feature Selection

The dimensionality of data is increasing very rapidly,which creates challenges for most of the current mining and learning algorithms,such as large memory requirements and high computational costs.The literature includes much research on feature selection for supervised learning.However,feature selection for unsupervised learning has only recently been studied.Finding the subset of features in unsupervised learning that enhances the performance is challenging since the clusters are indeterminate.This work proposes a hybrid technique for unsupervised feature selection called GAk-MEANS,which combines the genetic algorithm(GA)approach with the classical k-Means algorithm.In the proposed algorithm,a new fitness func-tion is designed in addition to new smart crossover and mutation operators.The effectiveness of this algorithm is demonstrated on various datasets.Fur-thermore,the performance of GAk-MEANS has been compared with other genetic algorithms,such as the genetic algorithm using the Sammon Error Function and the genetic algorithm using the Sum of Squared Error Function.Additionally,the performance of GAk-MEANS is compared with the state-of-the-art statistical unsupervised feature selection techniques.Experimental results show that GAk-MEANS consistently selects subsets of features that result in better classification accuracy compared to others.In particular,GAk-MEANS is able to significantly reduce the size of the subset of selected features by an average of 86.35%(72%–96.14%),which leads to an increase of the accuracy by an average of 3.78%(1.05%–6.32%)compared to using all features.When compared with the genetic algorithm using the Sammon Error Function,GAk-MEANS is able to reduce the size of the subset of selected features by 41.29%on average,improve the accuracy by 5.37%,and reduce the time by 70.71%.When compared with the genetic algorithm using the Sum of Squared Error Function,GAk-MEANS on average is able to reduce the size of the subset of selected features by 15.91%,and improve the accuracy by 9.81%,but the time is increased by a factor of 3.When compared with the machine-learning based methods,we observed that GAk-MEANS is able to increase the accuracy by 13.67%on average with an 88.76%average increase in time.

Different Feature Selection of Soil Attributes Influenced Clustering Performance on Soil Datasets

Feature selection is very important to obtain meaningful and interpretive clustering results from a clustering analysis. In the application of soil data clustering, there is a lack of good understanding of the response of clustering performance to different features subsets. In the present paper, we analyzed the performance differences between k-means, fuzzy c-means, and spectral clustering algorithms in the conditions of different feature subsets of soil data sets. The experimental results demonstrated that the performances of spectral clustering algorithm were generally better than those of k-means and fuzzy c-means with different features subsets. The feature subsets containing environmental attributes helped to improve clustering performances better than those having spatial attributes and produced more accurate and meaningful clustering results. Our results demonstrated that combination of spectral clustering algorithm with the feature subsets containing environmental attributes rather than spatial attributes may be a better choice in applications of soil data clustering.

Evaluation of Feature Subset Selection, Feature Weighting, and Prototype Selection for Biomedical Applications

Many medical diagnosis applications are characterized by datasets that contain under-represented classes due to the fact that the disease is much rarer than the normal case. In such a situation classifiers such as decision trees and Na?ve Bayesian that generalize over the data are not the proper choice as classification methods. Case-based classifiers that can work on the samples seen so far are more appropriate for such a task. We propose to calculate the contingency table and class specific evaluation measures despite the overall accuracy for evaluation purposes of classifiers for these specific data characteristics. We evaluate the different options of our case-based classifier and compare the perform-ance to decision trees and Na?ve Bayesian. Finally, we give an outlook for further work.

Apple leaf disease identification using genetic algorithm and correlation based feature selection method

Apple leaf disease is one of the main factors to constrain the apple production and quality.It takes a long time to detect the diseases by using the traditional diagnostic approach,thus farmers often miss the best time to prevent and treat the diseases.Apple leaf disease recognition based on leaf image is an essential research topic in the field of computer vision,where the key task is to find an effective way to represent the diseased leaf images.In this research,based on image processing techniques and pattern recognition methods,an apple leaf disease recognition method was proposed.A color transformation structure for the input RGB(Red,Green and Blue)image was designed firstly and then RGB model was converted to HSI(Hue,Saturation and Intensity),YUV and gray models.The background was removed based on a specific threshold value,and then the disease spot image was segmented with region growing algorithm(RGA).Thirty-eight classifying features of color,texture and shape were extracted from each spot image.To reduce the dimensionality of the feature space and improve the accuracy of the apple leaf disease identification,the most valuable features were selected by combining genetic algorithm(GA)and correlation based feature selection(CFS).Finally,the diseases were recognized by SVM classifier.In the proposed method,the selected feature subset was globally optimum.The experimental results of more than 90%correct identification rate on the apple diseased leaf image database which contains 90 disease images for there kinds of apple leaf diseases,powdery mildew,mosaic and rust,demonstrate that the proposed method is feasible and effective.

Person-specific named entity recognition using SVM with rich feature sets

Purpose: The purpose of the study is to explore the potential use of nature language process(NLP) and machine learning(ML) techniques and intents to find a feasible strategy and effective approach to fulfill the NER task for Web oriented person-specific information extraction.Design/methodology/approach: An SVM-based multi-classification approach combined with a set of rich NLP features derived from state-of-the-art NLP techniques has been proposed to fulfill the NER task. A group of experiments has been designed to investigate the influence of various NLP-based features to the performance of the system,especially the semantic features. Optimal parameter settings regarding with SVM models,including kernel functions,margin parameter of SVM model and the context window size,have been explored through experiments as well.Findings: The SVM-based multi-classification approach has been proved to be effective for the NER task. This work shows that NLP-based features are of great importance in datadriven NE recognition,particularly the semantic features. The study indicates that higher order kernel function may not be desirable for the specific classification problem in practical application. The simple linear-kernel SVM model performed better in this case. Moreover,the modified SVM models with uneven margin parameter are more common and flexible,which have been proved to solve the imbalanced data problem better.Research limitations/implications: The SVM-based approach for NER problem is only proved to be effective on limited experiment data. Further research need to be conducted on the large batch of real Web data. In addition,the performance of the NER system need be tested when incorporated into a complete IE framework.Originality/value: The specially designed experiments make it feasible to fully explore the characters of the data and obtain the optimal parameter settings for the NER task,leading to a preferable rate in recall,precision and F1measures. The overall system performance(F1value) for all types of name entities can achieve above 88.6%,which can meet the requirements for the practical application.

基于机器学习的镍基单晶高温合金蠕变寿命预测模型研究

构建合适的镍基单晶合金蠕变寿命预测模型,对于我国航空发动机叶片设计、强度分析和寿命预测具有重要意义。采用多项式回归、最近邻回归、支持向量机回归、决策树回归四种机器学习算法,建立镍基单晶高温合金蠕变寿命与合金成分、微观组织和蠕变工艺参数的关系模型,为镍基单晶高温合金的蠕变性能调控提供了新方法。基于蠕变寿命预测模型,系统地比较了四种算法和特征选择对模型性能的影响。结果表明,支持向量机回归模型的预测结果最优,相关性较高的四个特征依次为γ′固溶温度、Ta、W、Re。研究结果可为获得更有效的镍基单晶高温合金蠕变性能预测方法提供参考。

不均衡小样本下多特征优化选择的生命体触电故障识别方法

针对现有的剩余电流保护装置无法有效识别触电事故的问题,该文提出了一种不均衡小样本下多特征优化选择的生命体触电故障识别方法。首先通过变分自编码器(VAE)对实验收集到的生命体触电小样本数据进行增殖以实现正负样本均衡;然后在时域上提取能够反映波形动态变化特性的23个特征量,并利用高斯核Fisher判别分析(GKFDA)与最大信息系数(MIC)法从中选择最优表达特征组;最后,提出基于遗忘因子的在线顺序极限学习机(FOS-ELM)算法实现生命体触电行为的鉴别。实验结果表明,所提方法利用不均衡小样本触电数据集就可以训练出一个优秀的分类模型,诊断准确率可达98.75%,诊断时间仅为1.33 ms。其优良的性能结合在线增量式学习分类器设计,使得模型具备新知识学习能力,具有极好的工程应用前景。

基于mRMR-SOM的异步电机轴承故障诊断研究

针对异步电机轴承故障诊断问题,提出了一种融合最大相关最小冗余特征选择算法(mRMR)和自组织映射神经网络(SOM)的故障诊断方法,并将其应用于轴承故障诊断的不同阶段。首先,在实验室环境下搭建了异步电机故障诊断试验平台,在不同电机状态下分别采集振动、电流和电压信号,利用统计学方法获取了高维混合特征集;然后,以互信息为背景,利用mRMR根据特征与状态标签间的相关性和特征间的冗余性,筛选了具备强区分能力的特征,以避免计算冗余和后验诊断性能下降;最后,采用SOM对异步电机健康和轴承故障状态进行了分类识别,验证了SOM对异步电机轴承故障诊断的有效性,以及mRMR对故障诊断结果的影响。研究结果表明:基于mRMR-SOM的异步电机轴承故障诊断方法能够准确地区分健康和故障状态,测试集分类准确率达到89%;使用mRMR特征筛选能够将154维特征降低至17维,缩短23.5%的网络收敛时间,并将分类准确率由89%提升至98%;试验结果验证了基于mRMR-SOM的异步电机轴承故障诊断方法对于异步电机轴承故障诊断问题的有效性,且证实其具备良好的诊断效果。

面向高维不平衡医学数据的特征选择算法

基于传统机器学习分类算法对影像组学的高维不平衡数据分类结果不理想的问题,本文提出一种改进海洋捕食者的不平衡特征选择算法.首先,对海洋捕食者算法MPA算法进行改进,引入精英反向矩阵增加算法迭代后期的种群多样性,引入新的CF参数改善算法的收敛速度与精度,同时合理分配原始参数分布和取值来满足算法在不同阶段的搜索需求;接着针对不平衡数据引入新的目标函数来帮助MPA算法收敛到更优的特征子集.最后,基于G-means的精英反向海洋捕食者算法GEMPA算法在14个基础测试函数上进行测试并在12个公开数据集上与MPA,基于K个最近邻相关性的在线特征选择算法K-OFSD以及其余的6种元启发式算法GA、PSO、CSO、SSA、SCA和MFO对比分析.以平均F-measure值,平均特征数量,平均运行时间为评估指标,通过实验可知GEMPA算法能够快速搜索到分类精度最高的特征子集,降低高维数据的冗余度,针对改善高维不平衡数据分类问题有很好的发展潜力.

基于模型的非凸聚类算法

由于数据可能分布在非规则的流形上,其中潜在的簇往往呈现非凸的形状和结构,针对这类数据的聚类问题被统称为非凸聚类。现有的主流非凸聚类方法包括基于原始空间的方法和基于空间变换的方法,均忽略了非凸数据模式的显式描述。提出一种描述性模型用于非凸聚类。首先,基于核密度方法定义了一种具有混合形式的特征加权核密度模型,其无需事先假定任何概率分布模型且不限制簇的形状,这是传统基于模型的聚类方法无法实现的。其次,基于提出的模型推导了聚类目标函数,并基于期望最大化算法提出一种求解密度函数局部区域密度极大值的优化算法,那些上升到密度函数相同密度极大值的样本点被划分为同一个簇。最后,定义了一种基于模型的非凸聚类算法。算法不需人为定义簇的数量,并且能够为每个簇分配一个显式的概率密度函数,有助于更稳健和更准确地表征集群。除此之外,算法不仅在优化过程中进行自适应带宽选择,而且在优化过程中赋予了样本空间特征权重,实现了嵌入式特征选择。

适用于多种监督模型的特征选择方法研究

特征选择是模式识别、机器学习、数据挖掘等领域的重要问题之一,近年来已成为研究热点,并涌现出大量的用于选择特征的算法.现有的特征选择算法大多仅面向某一特定领域,其适用范围有限.采用基于Hilbert-Schmidt相关性标准的核方法衡量特征子集与目标对象间的相关程度,提出了一个适用性更广的特征选择方法FSM-HSIC,能较好地统一有监督、半监督和无监督3种模型下的特征选择过程,而且可从核方法的角度对整个过程进行抽象地描述,并深入理解现有的一些算法.同时以该方法为基础针对交互特征选择问题设计了新颖的FSI算法.理论分析和大量真实与仿真实验结果表明,与若干特征选择算法相比较,提出的算法具有良好的效率和稳定性,FSM-HSIC方法对新算法的产生具有重要的指导意义.

模拟电路故障特征提取的小波基选取方法研究

小波技术在高维故障特征数据的压缩及敏感信号提取已被广泛应用,但小波基的选取没有一个统一的标准;通过实际采样信号数据的小波分解、特征向量计算、波动性函数比较等技术对小波基函数的选取进行了研究;最后通过综合小波分析、神经网络等技术的模拟电路故障诊断系统的诊断实例验证了所提选取方法的有效性;使用9种常用小波基函数,分别对采样信号进行分解并计算波动性函数,并在模拟电路故障诊断系统进行验证;小波基函数bior2.2的波动较小且与诊断结果一致。

基于多时相遥感影像的北京平原人工林树种分类

为解决传统遥感分类方法区分平原人工造林地树种难度较大的问题,利用4个不同时相的高空间分辨率卫星影像,基于ESP计算方差变化率并结合目视解译获取影像的最佳分割尺度;通过相关系数法筛选构建的特征,进行面向对象的多时相影像和单时相影像分类,并与基于像元分类方法进行对比分析.结果表明:基于多时相影像各类别分类精度为64%,高于单时相分类精度(51%);面向对象KNN方法的分类精度优于SVM和MLC分类方法,两者精度分别为49%和43%.在树种丰富且分布复杂的平原造林林地景观中,利用多时相遥感数据,采用面向对象分类方法用于树种精细分类更具优势.

基于稀疏组lasso的脑机接口通道和特征选择研究

脑电信号(EEG)特征提取和分类是脑机接口(BCI)系统的核心问题之一。由于BCI系统中EEG信号多通道采样和特征向量的高维性,有效的特征选择算法已经成为研究中不可分割的一部分。针对EEG特征选择问题采用一种新方法:基于封装式稀疏组lasso的EEG融合特征的同时通道和特征选择方法。实验中将该方法与现有的通道选择和特征选择方法进行比较,结果表明,该方法更适用于高维融合特征的最优特征子集选择问题,且该算法稳定、时间成本低。此外,在保证错误率相当或较低的情况下,该方法能够同时实现通道和特征选择。国际BCI竞赛IV的两类运动想象信号的测试错误率为15.28%。

汉语句子谓语中心词的自动识别

谓语中心词的识别是句法成分分析中的一个非常重要的部分。本文提出了一种规则和特征学习相结合的谓语识别方法 ,将整个谓语识别的过程分为语片捆绑、谓语粗筛选和谓语精筛选三个阶段。在谓语粗筛选中 ,利用规则过滤掉明显不能充当谓语的词 ,得到一个准谓语集 ;在精筛选阶段 ,选择谓语的支持特征 ,根据统计计算得到每个特征对谓语的支持度 ,然后利用准谓语在句子中的上下文出现的特征对准谓语集中的词进行再次筛选 ,从而确定出句子的谓语中心词。经过测试表明 。

一种新的基于统计的自动文本分类方法

自动文本分类就是在给定的分类体系下 ,让计算机根据文本的内容确定与它相关联的类别。为了提高分类性能 ,本文提出了中文文本多层次特征提取方法和基于核的距离加权KNN算法。多层次特征提取方法在汉字、常用词表和专业词表三个层次上提取文档的统计特征 ,能够更好地反映文档的统计分布。基于核的距离加权KNN算法解决了样本的多峰分布、边界重叠问题和分类器的精确分类决策问题。实际应用中 ,互联网和文本库提供了大量经过粗分类的训练文本 ,但普遍存在样本质量较差的问题 ,本文通过样本重要性分析技术解决此问题。实验系统证明了新方法的有效性。

一种基于粒子群优化的目标跟踪特征选择算法

针对复杂背景下的运动目标跟踪特征选择问题,提出了一种基于粒子群优化的目标跟踪特征选择算法。假设具有目标与背景间最好可分离性的特征为最好的跟踪特征。通过构建目标与背景的图像特征分布方差的比值函数作为衡量目标与背景间的可分离性判据。使用粒子群优化算法优化不同的特征组合实时获取最优的目标跟踪特征。为验证该算法的有效性,将选择的最优特征与一种基于核的跟踪算法相结合进行跟踪实验。实验结果表明,算法能有效提高传统基于核的跟踪算法对于复杂场景下的运动目标跟踪的鲁棒性与准确性。