Binarization of degraded document image based on feature space partitioning and classification

In this paper, we propose a new algorithm for the binarization of degraded document images. We map the image into a 2D feature space in which the text and background pixels are separable, and then we partition this feature space into small regions. These regions are labeled as text or background using the result of a basic binarization algorithm applied on the original image. Finally, each pixel of the image is classified as either text or background based on the label of its corresponding region in the feature space. Our algorithm splits the feature space into text and background regions without using any training dataset. In addition, this algorithm does not need any parameter setting by the user and is appropriate for various types of degraded document images. The proposed algorithm demonstrated superior performance against six well-known algorithms on three datasets.     

Learning model order from labeled and unlabeled data for partially supervised classification, with application to word sense disambiguation

Previous partially supervised classification methods can partition unlabeled data into positive examples and negative examples for a given class by learning from positive labeled examples and unlabeled examples, but they cannot further group the negative examples into meaningful clusters even if there are many different classes in the negative examples. Here we proposed an automatic method to obtain a natural partitioning of mixed data (labeled data + unlabeled data) by maximizing a stability criterion defined on classification results from an extended label propagation algorithm over all the possible values of model order (or the number of classes) in mixed data. Our experimental results on benchmark corpora for word sense disambiguation task indicate that this model order identification algorithm with the extended label propagation algorithm as the base classifier outperforms SVM, a one-class partially supervised classification algorithm, and the model order identification algorithm with semi-supervised k-means clustering as the base classifier when labeled data is incomplete.     

Research of fast SOM clustering for text information

The state-of-the-art text clustering methods suffer from the huge size of documents with high-dimensional features. In this paper, we studied fast SOM clustering technology for Text Information. Our focus is on how to enhance the efficiency of text clustering system whereas high clustering qualities are also kept. To achieve this goal, we separate the system into two stages: offline and online. In order to make text clustering system more efficient, feature extraction and semantic quantization are done offline. Although neurons are represented as numerical vectors in high-dimension space, documents are represented as collections of some important keywords, which is different from many related works, thus the requirement for both time and space in the offline stage can be alleviated. Based on this scenario, fast clustering techniques for online stage are proposed including how to project documents onto output layers in SOM, fast similarity computation method and the scheme of Incremental clustering technology for real-time processing, We tested the system using different datasets, the practical performance demonstrate that our approach has been shown to be much superior in clustering efficiency whereas the clustering quality are comparable to traditional methods.     

基于朴素贝叶斯和遗传算法的两类文本分类方法

本文提出了一种基于朴素贝叶斯和遗传算法的两类文本分类方法,该方法将朴素贝叶斯分类器变换为在二维空间中的一条分割线,在分割线临近的文本分类不可靠区间内,利用遗传算法搜索最优文本分割线,从而使分类器达到最佳性能.在由12600篇文本构成的中文语料数据集上的实验表明,该方法具有较高的分类性能和效率,查准率、查全率和F1值分别达到97.98%,91.05%和94.39%.     

一种新的基于聚类的多分类器融合算法

提出了一种新的多分类器融合算法,该算法能找出各分类器在特征空间中局部性能较好的区域,并利用具有最优局部性能的分类器的输出作为最终的融合结果。首先,利用各分类器对训练样本进行分类,这样训练样本被划分为正确分类样本和错误分类样本两个集合;接着,对这两个样本集合分别进行聚类分析来划分特征空间,并计算各分类器在特征空间局部区域中的性能;在测试时,选择测试样本周围局部性能最优的分类器的输出作为最终的融合结果。基于ELENA数据集的实验显示了该算法的有效性。     

Sample re-weighting hyper box classifier for multi-class data classification

In this work, we propose two novel classifiers for multi-class classification problems using mathematical programming optimisation techniques. A hyper box-based classifier (Xu & Papageorgiou, 2009) that iteratively constructs hyper boxes to enclose samples of different classes has been adopted. We firstly propose a new solution procedure that updates the sample weights during each iteration, which tweaks the model to favour those difficult samples in the next iteration and therefore achieves a better final solution. Through a number of real world data classification problems, we demonstrate that the proposed refined classifier results in consistently good classification performance, outperforming the original hyper box classifier and a number of other state-of-the-art classifiers.Furthermore, we introduce a simple data space partition method to reduce the computational cost of the proposed sample re-weighting hyper box classifier. The partition method partitions the original dataset into two disjoint regions, followed by training sample re-weighting hyper box classifier for each region respectively. Through some real world datasets, we demonstrate the data space partition method considerably reduces the computational cost while maintaining the level of prediction accuracies.     

Exponentially weighted moving average charts for detecting concept drift

Classifying streaming data requires the development of methods which are computationally efficient and able to cope with changes in the underlying distribution of the stream, a phenomenon known in the literature as concept drift. We propose a new method for detecting concept drift which uses an exponentially weighted moving average (EWMA) chart to monitor the misclassification rate of an streaming classifier. Our approach is modular and can hence be run in parallel with any underlying classifier to provide an additional layer of concept drift detection. Moreover our method is computationally efficient with overhead O(1) and works in a fully online manner with no need to store data points in memory. Unlike many existing approaches to concept drift detection, our method allows the rate of false positive detections to be controlled and kept constant over time.     

Sentiment visualization and classification via semi-supervised nonlinear dimensionality reduction

Sentiment analysis, which detects the subjectivity or polarity of documents, is one of the fundamental tasks in text data analytics. Recently, the number of documents available online and offline is increasing dramatically, and preprocessed text data have more features. This development makes analysis more complex to be analyzed effectively. This paper proposes a novel semi-supervised Laplacian eigenmap (SS-LE). The SS-LE removes redundant features effectively by decreasing detection errors of sentiments. Moreover, it enables visualization of documents in perceptible low dimensional embedded space to provide a useful tool for text analytics. The proposed method is evaluated using multi-domain review data set in sentiment visualization and classification by comparing other dimensionality reduction methods. SS-LE provides a better similarity measure in the visualization result by separating positive and negative documents properly. Sentiment classification models trained over reduced data by SS-LE show higher accuracy. Overall, experimental results suggest that SS-LE has the potential to be used to visualize documents for the ease of analysis and to train a predictive model in sentiment analysis. SS-LE can also be applied to any other partially annotated text data sets.     

情感分类研究进展

文本情感分析是多媒体智能理解的重要问题之一.情感分类是情感分析领域的核心问题,旨在解决评论情感极性的自动判断问题.由于互联网评论数据规模与日俱增,传统基于词典的方法和基于机器学习的方法已经不能很好地处理海量评论的情感分类问题.随着近年来深度学习技术的快速发展,其在大规模文本数据的智能理解上表现出了独特的优势,越来越多的研究人员青睐于使用深度学习技术来解决文本分类问题.主要分为2个部分：1)归纳总结传统情感分类技术,包括基于字典的方法、基于机器学习的方法、两者混合方法、基于弱标注信息的方法以及基于深度学习的方法;2)针对前人情感分类方法的不足,详细介绍所提出的面向情感分类问题的弱监督深度学习框架.此外,还介绍了评论主题提取相关的经典工作.最后,总结了情感分类问题的难点和挑战,并对未来的研究工作进行了展望.     

基于BERT的社交电商文本分类算法

随着网络购物的高速发展,网络商家和购物者在网络交易活动中产生了大量的交易数据,其中蕴含着巨大的分析价值。针对社交电商商品文本的文本分类问题,为了更加高效准确地判断文本所描述商品的类别,提出了一种基于BERT模型的社交电商文本分类算法。首先,该算法采用BERT(Bidirectional Encoder Representations from Transformers)预训练语言模型来完成社交电商文本的句子层面的特征向量表示,随后有针对性地将获得的特征向量输入分类器进行分类,最后采用社交电商文本的数据集进行算法验证。实验结果表明,经过训练的模型在测试集上的分类结果F1值最高可达94.61%,高出BERT模型针对MRPC的分类任务6%。因此,所提社交电商文本分类算法能够较为高效准确地判断文本所描述商品的类别,有助于进一步分析网络交易数据,从海量数据中提取有价值的信息。     

基于卡方特征选择和LDA主题模型的中文短文本分类

中文短文本自身包含词汇个数少、描述信息能力弱,常用的文本分类方法对于短文本分类效果不理想。同时传统的文本分类方法在处理大规模文本分类时会出现向量维数很高的情况,造成算法效率低,而且一般用于长文本分类的特征选择方法都是基于数理统计的,忽略了文本中词项之间的语义关系。针对以上问题本文提出基于卡方特征选择和LDA主题模型的中文短文本分类方法,方法使用LDA主题模型的训练结果对传统特征选择方法进行特征扩展,以达到将数理信息和语义信息融入分类算法的目的。对比试验表明,这种方法提高了中文短文本分类效果。     

Automated classification of building information modeling (BIM) case studies by BIM use based on natural language processing (NLP) and unsupervised learning

This paper comparatively analyzes a method to automatically classify case studies of building information modeling (BIM) in construction projects by BIM use. It generally takes a minimum of thirty minutes to hours of collection and review and an average of four information sources to identify a project that has used BIM in a manner that is of interest. To automate and expedite the analysis tasks, this study deployed natural language processing (NLP) and commonly used unsupervised learning for text classification, namely latent semantic analysis (LSA) and latent Dirichlet allocation (LDA). The results were validated against one of representative supervised learning methods for text classification—support vector machine (SVM). When LSA and LDA detected phrases in a BIM case study that had higher similarity values to the definition of each BIM use than the threshold values, the system determined that the project had deployed BIM in the detected approach. For the classification of BIM use, the BIM uses specified by Pennsylvania State University were utilized. The approach was validated using 240 BIM case studies (512,892 features). When BIM uses were employed in a project, the project was labeled as “1”; when they were not, the project was labeled as “0.” The performance was analyzed by changing parameters: namely, document segmentation, feature weighting, dimensionality reduction coefficient (k-value), the number of topics, and the number of iterations. LDA yielded the highest F1 score, 80.75% on average. LDA and LSA yielded high recall and low precision in most cases. Conversely, SVM yielded high precision and low recall in most cases and fluctuations in F1 scores.     

基于赋权粗糙隶属度的文本情感分类方法

提出了基于赋权粗糙隶属度的文本情感分类方法.该方法将特征倾向强度引入到文本的向量空间表示法中,建立了基于二元组属性(特征,特征倾向强度)的文本表示模型.提出了基于情感倾向强度序的属性离散化方法,将特征选择寓于离散化过程,达到数据降维的目的.利用特征倾向强度,定义了赋权粗糙隶属度,用于新文本的情感分类.在真实汽车评论语料上,与支持向量机分类模型进行比较实验表明,基于赋权粗糙隶属度的文本情感分类方法在对数据进行一定程度的压缩后仍表现出较好的分类性能.     

基于上下文重构的短文本情感极性判别研究

文本对象所固有的多义性,面对短文本特征稀疏和上下文缺失的情况,现有处理方法无法明辨语义,形成了底层特征和高层表达之间巨大的语义鸿沟.本文尝试借由时间、空间、联系等要素挖掘文本间隐含的关联关系,重构文本上下文范畴,提升情感极性分类性能.具体做法对应一个两阶段处理过程:1)基于短文本的内在联系将其初步重组成上下文(领域);2)将待处理短文本归入适合的上下文(领域)进行深入处理.首先给出了基于Naive Bayes分类器的短文本情感极性分类基本框架,揭示出上下文(领域)范畴差异对分类性能的影响.接下来讨论了基于领域归属划分的文本情感极性分类增强方法,并将领域的概念扩展为上下文关系,提出了基于特殊上下文关系的文本情感极性判别方法.同时为了解决由于信息缺失所造成的上下文重组困难,给出基于遗传算法的任意上下文重组方案.理论分析表明,满足限制条件的前提下,基于上下文重构的情感极性判别方法能够同时降低抽样误差(Sample error)和近似误差(Approximation error).真实数据集上的实验结果也验证了理论分析的结论.     

FSKNN: Multi-label text categorization based on fuzzy similarity and k nearest neighbors

We propose an efficient approach, FSKNN, which employs fuzzy similarity measure (FSM) and k nearest neighbors (KNN), for multi-label text classification. One of the problems associated with KNN-like approaches is its demanding computational cost in finding the k nearest neighbors from all the training patterns. For FSKNN, FSM is used to group the training patterns into clusters. Then only the training documents in those clusters whose fuzzy similarities to the document exceed a predesignated threshold are considered in finding the k nearest neighbors for the document. An unseen document is labeled based on its k nearest neighbors using the maximum a posteriori estimate. Experimental results show that our proposed method can work more effectively than other methods.     

Extending boosting for large scale spoken language understanding

We propose three methods for extending the Boosting family of classifiers motivated by the real-life problems we have encountered. First, we propose a semisupervised learning method for exploiting the unlabeled data in Boosting. We then present a novel classification model adaptation method. The goal of adaptation is optimizing an existing model for a new target application, which is similar to the previous one but may have different classes or class distributions. Finally, we present an efficient and effective cost-sensitive classification method that extends Boosting to allow for weighted classes. We evaluated these methods for call classification in the AT&;T VoiceTone^® spoken language understanding system. Our results indicate that it is possible to obtain the same classification performance by using 30% less labeled data when the unlabeled data is utilized through semisupervised learning. Using model adaptation we can achieve the same classification accuracy using less than half of the labeled data from the new application. Finally, we present significant improvements in the “important” (i.e., higher weighted) classes without a significant loss in overall performance using the proposed cost-sensitive classification method.     

基于优化样本分布抽样集成学习的半监督文本分类方法研究

针对现有文本分类方法在即时性文本信息上面临的挑战,考虑到即时性文本信息具有已标注数据规模小的特点,为了提高半监督学习的分类性能,该文提出一种基于优化样本分布抽样集成学习的半监督文本分类方法。首先,通过运用一种新的样本抽样的优化策略,获取多个新的子分类器训练集,以增加训练集之间的多样性和减少噪声的扩散范围,从而提高分类器的总体泛化能力;然后,采用基于置信度相乘的投票机制对预测结果进行集成,对未标注数据进行标注;最后,选取适量的数据来更新训练模型。实验结果表明,该方法在长文本和短文本上都取得了优于研究进展方法的分类性能。     

Dissimilarity-based classification in the absence of local ground truth: Application to the diagnostic interpretation of chest radiographs

In this paper classification on dissimilarity representations is applied to medical imaging data with the task of discrimination between normal images and images with signs of disease. We show that dissimilarity-based classification is a beneficial approach in dealing with weakly labeled data, i.e. when the location of disease in an image is unknown and therefore local feature-based classifiers cannot be trained. A modification to the standard dissimilarity-based approach is proposed that makes a dissimilarity measure multi-valued, hence, able to retain more information. A multi-valued dissimilarity between an image and a prototype becomes an image representation vector in classification. Several classification outputs with respect to different prototypes are further integrated into a final image decision. Both standard and proposed methods are evaluated on data sets of chest radiographs with textural abnormalities and compared to several feature-based region classification approaches applied to the same data. On a tuberculosis data set the multi-valued dissimilarity-based classification performs as well as the best region classification method applied to the fully labeled data, with an area under the receiver operating characteristic (ROC) curve (A_z) of 0.82. The standard dissimilarity-based classification yields A_z=0.80. On a data set with interstitial abnormalities both dissimilarity-based approaches achieve A_z=0.98 which is closely behind the best region classification method.     

一种利用近邻和信息熵的主动文本标注方法

由于大规模标注文本数据费时费力,利用少量标注样本和大量未标注样本的半监督文本分类发展迅速.在半监督文本分类中,少量标注样本主要用来初始化分类模型,其合理性将影响最终分类模型的性能.为了使标注样本尽可能吻合原始数据的分布,提出一种避开选择已标注样本的K近邻来抽取下一组候选标注样本的方法,使得分布在不同区域的样本有更多的标注机会.在此基础上,为了获得更多的类别信息,在候选标注样本中选择信息熵最大的样本作为最终的标注样本.真实文本数据上的实验表明了提出方法的有效性.     

Text feature selection for sentiment classification of Chinese online reviews

In order to meet the requirement of customised services for online communities, sentiment classification of online reviews has been applied to study the unstructured reviews so as to identify users’ opinions on certain products. The purpose of this article is to select features for sentiment classification of Chinese online reviews with techniques well performed in traditional text classification. First, adjectives, adverbs and verbs are identified as the potential text features containing sentiment information. Then, four statistical feature selection methods, such as document frequency (DF), information gain (IG), chi-squared statistic (CHI) and mutual information (MI), are adopted to select features. After that, the Boolean weighting method is applied to set feature weights and construct a vector space model. Finally, a support vector machine (SVM) classifier is employed to predict the sentiment polarity of online reviews. Comparative experiments are conducted based on hotel online reviews in Chinese. The results indicate that the highest accuracy of the sentiment classification of Chinese online reviews is achieved by taking adjectives, adverbs and verbs together as the feature. Besides that, different feature selection methods make distinct performances on sentiment classification, as DF performs the best, CHI follows and IG ranks the last, whereas MI is not suitable for sentiment classification of Chinese online reviews. This conclusion will be helpful to improve the accuracy of sentiment classification and be useful for further research.