With the dramatically development of Internet, the information processing and management technology onWWW have become a great important branch of data mining and data warehouse. Especially, nowadays, Text Miningis mar...With the dramatically development of Internet, the information processing and management technology onWWW have become a great important branch of data mining and data warehouse. Especially, nowadays, Text Miningis marvelously emerging and plays an important role in interrelated fields. So it is worth summarizing the contentabout text mining from its definition to relational methods and techniques. In this paper, combined to comparativelymature data mining technology, we present the definition of text mining and the multi-stage text mining process mod-el. Moreover, this paper roundly introduces the key areas of text mining and some of the powerful text analysis tech-niques, including: Word Automatic Segmenting, Feature Representation, Feature Extraction, Text Categorization,Text Clustering, Text Summarization, Information Extraction, Pattern Quality Evaluation, etc. These techniquescover the whole process from information preprocessing to knowledge obtaining.展开更多
高光谱图像目标检测作为一个研究热点在军事和民用方面的应用越来越广泛。为了能同时利用高光谱图像数据的空谱信息,本文提出一种新的基于张量表示的高光谱图像目标检测算法。算法使用CP(Canonical Polyadic)张量分解技术和张量块分解(B...高光谱图像目标检测作为一个研究热点在军事和民用方面的应用越来越广泛。为了能同时利用高光谱图像数据的空谱信息,本文提出一种新的基于张量表示的高光谱图像目标检测算法。算法使用CP(Canonical Polyadic)张量分解技术和张量块分解(Block Term Decomposition,BTD)分别对高光谱数据进行盲源分析,提取了有效的局部图像块空谱特征,建立了一个基于稀疏表示和协作表示的检测模型,针对多种类型背景复杂的场景数据进行实验,并与当前流行的目标检测算法进行比较。从可视化检测结果来看,本文算法在复杂背景和强噪声环境下,有效提取了空谱特征,对背景具有较好的抑制能力,检测的目标显著。此外,本文从接收机操作曲线(Receiver Operating Characteristic Curve,ROC)和ROC曲线下面积(Area Under Curve,AUC)等定量指标分析算法性能。以较为流行的Sandiego图像为例,在10%的虚警率下,本文算法取得90%的检测精度,AUC大于0.95。本文算法相较几种流行算法而言具有较高的检测精度,更强的鲁棒性。展开更多
文摘With the dramatically development of Internet, the information processing and management technology onWWW have become a great important branch of data mining and data warehouse. Especially, nowadays, Text Miningis marvelously emerging and plays an important role in interrelated fields. So it is worth summarizing the contentabout text mining from its definition to relational methods and techniques. In this paper, combined to comparativelymature data mining technology, we present the definition of text mining and the multi-stage text mining process mod-el. Moreover, this paper roundly introduces the key areas of text mining and some of the powerful text analysis tech-niques, including: Word Automatic Segmenting, Feature Representation, Feature Extraction, Text Categorization,Text Clustering, Text Summarization, Information Extraction, Pattern Quality Evaluation, etc. These techniquescover the whole process from information preprocessing to knowledge obtaining.
文摘高光谱图像目标检测作为一个研究热点在军事和民用方面的应用越来越广泛。为了能同时利用高光谱图像数据的空谱信息,本文提出一种新的基于张量表示的高光谱图像目标检测算法。算法使用CP(Canonical Polyadic)张量分解技术和张量块分解(Block Term Decomposition,BTD)分别对高光谱数据进行盲源分析,提取了有效的局部图像块空谱特征,建立了一个基于稀疏表示和协作表示的检测模型,针对多种类型背景复杂的场景数据进行实验,并与当前流行的目标检测算法进行比较。从可视化检测结果来看,本文算法在复杂背景和强噪声环境下,有效提取了空谱特征,对背景具有较好的抑制能力,检测的目标显著。此外,本文从接收机操作曲线(Receiver Operating Characteristic Curve,ROC)和ROC曲线下面积(Area Under Curve,AUC)等定量指标分析算法性能。以较为流行的Sandiego图像为例,在10%的虚警率下,本文算法取得90%的检测精度,AUC大于0.95。本文算法相较几种流行算法而言具有较高的检测精度,更强的鲁棒性。
