Mixture of KL subspaces for relevance feedback

Relevance feedback has recently emerged as a solution to the problem of improving the retrieval performance of an image retrieval system based on low-level information such as color, texture and shape features. Most of the relevance feedback approaches limit the utilization of the user’s feedback to a single search session, performing a short-term learning. In this paper we present a novel approach for short and long term learning, based on the definition of an adaptive similarity metric and of a high level representation of the images. For short-term learning, the relevant and non-relevant information given by the user during the feedback process is employed to create a positive and a negative subspace of the feature space. For long-term learning, the feedback history of all the users is exploited to create and update a representation of the images which is adopted for improving retrieval performance and progressively reducing the semantic gap between low-level features and high-level semantic concepts. The experimental results prove that the proposed method outperforms many other state of art methods in the short-term learning, and demonstrate the efficacy of the representation adopted for the long-term learning.

基于近似Markov Blanket和动态互信息的特征选择算法

针对大量无关和冗余特征的存在可能降低分类器性能的问题,提出了一种基于近似Markov Blanket和动态互信息的特征选择算法。该算法利用互信息作为特征相关性的度量准则,并在未识别的样本上对互信息进行动态估值,利用近似Markov Blanket原理准确地去除冗余特征,从而获得远小于原始特征规模的特征子集。通过仿真试验证明了该算法的有效性。以支持向量机为分类器,在公共数据集UCI上进行了试验,并与DMIFS和ReliefF算法进行了对比。试验结果证明,该算法选取的特征子集与原始特征子集相比,以远小于原始特征规模的特征子集获得了高于或接近于原始特征集合的分类结果。     

基于贝叶斯网络模型的用户兴趣联合推送

The association push of interesting information is implemented according to the information that other users with similar interests have read, the refore user can easily find high quality interesting information from the immense Web information. The association push of interesting information based on Bayesian network model ispresented in this paper, it stores the conditional probability and semantic meaning between the terms, so the similarity between the users' interests and the similarity between the user's interest and document can be computed according to the semantic meaning of the terms, therefore the association push of interesting information can beimplemented. The experience result indicates this method is effective.     

基于隐式反馈的个人信息检索技术及实现

回顾了已有的相关反馈技术，在此基础上提出了构造和调整用户兴趣模型的隐式反馈算法，给出了一个基于隐式反馈的lnfoAgent的设计实现和实验结果，实验表明隐式反馈技术对提高检索精度有很大的帮助。     

手绘草图识别中的相关反馈方法

手绘草图是人类思维外化和表达设计意图的有效工具之一,手绘草图的模糊性和用户适应性问题是草图识别中的关键问题。本文提出了将相关反馈机制引入到手绘草图识别中以捕捉用户意图的方法,该方法以抽取手绘草图的向量化特征为基础,首先利用基于图形特征的相似度计算,给出手绘草图候选识别结果集,然后借助用户对识别结果的相关性评价,通过逐渐调整图形构成特征的权重来捕捉用户输入意图,并提高识别效果。实验表明本文所提出的方法具有很好的效果。     

基于LE与ICROA-RVM的瓦斯传感器故障诊断

针对瓦斯传感器故障诊断速度慢、诊断精度不高的问题,以常见的冲击型、漂移型、偏置型和周期型传感器输出故障作为研究对象,提出一种基于拉普拉斯特征映射(LE)和改进化学反应优化算法(ICROA)优化的相关向量机(RVM)进行模式分类与辨识,实现瓦斯传感器故障诊断。首先采用流形学习方法 LE对高维原始数据空间进行非线性降维特征提取,提取故障特征,该方法极大地保留了原始数据中的整体几何信息;然后将故障特征作为RVM模型训练输入,利用ICROA算法对RVM模型的核参数进行全局寻优,将训练好的ICROA-RVM模型对测试样本进行故障诊断。实验结果表明:该诊断方法具有训练速度快,故障辨识精度高的特点,故障诊断正确率在96%以上,能够有效地提高瓦斯传感器故障诊断的速度和准确性。     

改进相关反馈技术在CBVR人体动作识别中的应用研究

针对CBVR人体动作识别问题,提出利用相关反馈技术与其它技术相结合。首先,分析了从视频中提取人类行为并将其表示成特征集的方法;然后研究了各种基于内容的视频检索方法;最后,将这些技术进行组合得到基于相关反馈技术的最优CBVR人体动作识别方法。在三个行为数据库中(包括UCF运动,UCF视频网站和HOHA2)评估了由上述几种技术组合而成的性能,实验结果为开发高效的基于内容的视频检索系统提供了有益的借鉴。     

Prediction of the Melt Index Based on the Relevance Vector Machine with Modified Particle Swarm Optimization

A novel chemical soft‐sensor approach for the prediction of the melt index (MI) in the propylene polymerization industry is presented. The MI is considered as one of the important variables of quality that determine the product specifications. Thus, a reliable estimation of the MI is crucial in quality control. An accurate optimal predictive model of MI values with the relevance vector machine (RVM) is proposed, where the RVM is employed to build the MI prediction model; a modified particle swarm optimization (MPSO) algorithm is then introduced to optimize the parameter of the RVM, and the MPSO‐RVM model is thereby developed. An online correcting strategy (OCS) is further carried out to update the modeling data and to revise the model's parameter self‐adaptively whenever model mismatch happens. Based on the data from a real polypropylene production plant, a detailed comparison is carried out among the least squares support vector machine (LS‐SVM), RVM, MPSO‐RVM, and OCS‐MPSO‐RVM models. The research results reveal the prediction accuracy and validity of the proposed approach.     

An adaptive technique for content-based image retrieval

We discuss an adaptive approach towards Content-Based Image Retrieval. It is based on the Ostensive Model of developing information needs—a special kind of relevance feedback model that learns from implicit user feedback and adds a temporal notion to relevance. The ostensive approach supports content-assisted browsing through visualising the interaction by adding user-selected images to a browsing path, which ends with a set of system recommendations. The suggestions are based on an adaptive query learning scheme, in which the query is learnt from previously selected images. Our approach is an adaptation of the original Ostensive Model based on textual features only, to include content-based features to characterise images. In the proposed scheme textual and colour features are combined using the Dempster-Shafer theory of evidence combination. Results from a user-centred, work-task oriented evaluation show that the ostensive interface is preferred over a traditional interface with manual query facilities. This is due to its ability to adapt to the user's need, its intuitiveness and the fluid way in which it operates. Studying and comparing the nature of the underlying information need, it emerges that our approach elicits changes in the user's need based on the interaction, and is successful in adapting the retrieval to match the changes. In addition, a preliminary study of the retrieval performance of the ostensive relevance feedback scheme shows that it can outperform a standard relevance feedback strategy in terms of image recall in category search.