位置：立项数据库 > 立项详情页
WEB搜索与挖掘的新理论与方法
  • 项目名称：WEB搜索与挖掘的新理论与方法
  • 项目类别：重点项目
  • 批准号：60933004
  • 申请代码：F0205
  • 项目来源：国家自然科学基金
  • 研究期限：2010-01-01-2013-12-31
  • 项目负责人：李晓明
  • 负责人职称：教授
  • 依托单位：北京大学
  • 批准年度：2009
中文摘要：

瞄准Web 3.0的智能搜索引擎，深入研究Web搜索与挖掘的新理论和方法（1）研究Web的性质与演化规律，探索Web信息的基本特征，提出Web信息搜集的新模式和方法；（2）研究Web信息的语义表示与推理，探索异构Web信息关联、集成与重构的逻辑基础，提出Web异构信息的语义模型；（3）研究Web数据的挖掘与组织，从结构、内容和用户行为诸方面探索Web信息的模式，针对Web信息的异构性和时态性，建立支持高效访问的数据组织方式；（4）针对Web信息海量性的特点，研究分布并行挖掘理论与方法，为潜在的实际应用提供高性能算法和支持环境；（5）研究图像视频的语义自动标注，综合利用概念语义与关联信息，提高对Web图像视频理解的能力，提出Web多模态检索的新方法。 在上述理论研究成果的基础上，研制Web智能搜索引擎原型系统，验证本项目研究中提出的新理论和新方法。

中文主题词： Web搜索；语义表示；并行挖掘；数据组织；
英文摘要：

Web Search；semantic representation；parallel mining；data organization；

英文主题词： Web Search；semantic representation；parallel mining；data organization；
结论摘要：

能够从海量Web信息中挖掘出有用的知识，不仅有学术上的重要意义（证明我们对Web信息的形式、结构、时态等有深刻的理解），而且还有潜在巨大的社会经济价值。其研究难点在于Web信息的多样性、随机性、随意性、非结构化，而搜索与挖掘的应用常常追求的是相对比较具体明确的结果，从而使得发现一般且有用的规律十分困难。基于深度学习技术，利用云计算环境，结合众包机制，是Web搜索与挖掘理论与方法发展的一个新趋势。本项目主要成果包括以下五个方面。 1.提出了一种对微博和在线短评论进行局部化主题建模方法。该方法以局部文档为单位代替以词为单位打标签的方法，同时引入了背景信息消除词汇噪音，对微博商品评论的情感分析效果很好，论文被Google引用170次。 2.提出利用多种特征关联信息的融合和动态交互构建对象之间的相似关系和时序关系，设计多种数据对象模型，索引检索Web数据对象，能有效支持热点探测、分类、检索等Web应用，论文发表在SIGMOD、SIGIR、AAAI、ICDE等CCF A类国际会议，被Google引用50次。 3. 提出了一种演进式新闻文档摘要的生成方法，在摘要迭代过程中综合考虑了全局信息和局部信息，在纯文本摘要，文本和图像混合摘要，诗词生成等方面效果很好。在SIGIR、IJCAI、JCDL（最佳学生论文提名）、CIKM上发表了长文，博士生严睿获得了2012北京大学五四奖章。 4.提出了基于一致性正则化的多源跨领域迁移学习框架，能利用源领域上的局部数据，而且考虑了子分类器在目标领域上的预测的一致性。还提出挖掘领域间共性与特性的协同PLSA模型，从生成模型角度解决了从多领域间挖掘共享概念的学习问题，上述工作在IEEETKDE发表两篇论文，Google他引70次，庄福振博士获得2013年人工智能学会优秀博士论文。 5.开发了“天网搜索－中国事件检索与发现”系统。该系统整合了1000多万篇网页、文本、视频、微博数据，体现了多源大跨度建模等多项技术。同时也开发了Web数据并行挖掘云服务平台，提供从数据采集到数据预处理再到文本挖掘的一站式服务，并推广到并推广到电信、信息安全等领域。

成果综合统计
成果类型
数量
  • 期刊论文
  • 会议论文
  • 专利
  • 获奖
  • 著作
  • 74
  • 88
  • 1
  • 0
  • 2
期刊论文
网页去噪:研究综述
A Framework for Similarity Search of Time Series Cliques with Natural Relations
CHSMST: A clustering algorithm based on hyper surface and minimum spanning tree
Approaches to Exploring Category Information for Question Retrieval in Community Question-Answer Arc
A feature binding computational model for multi-class object categorization and recognition
基于数据共性的鲁棒特征表示
Parallel extreme learning machine for regression based on MapReduce
一种基于特征整合理论的物体识别模型(英文)
Clustering in extreme learning machine feature space
问答系统研究综述
一种基于特征捆绑计算模型的物体识别方法
基于Agent的专家系统推理模型
Model for Rumor Propagation on Social Networks
A dynamic hot topic extraction model based on time window
D-LDA: A topic modeling approach without constraint generation for semi-defined classification
Modeling continuous visual features for semantic image annotation and retrieval
Multi-granularity context model for dynamic Web service composition
Fusing semantic aspects for image annotation and retrieval
基于动态描述逻辑的服务组合及质量模型
建模连续视觉特征的图像语义标注方法
一种基于语义轨迹的事件规则学习算法
融合语义主题的图像自动标注
基于进化规划的FCMBP模糊聚类改进方法
权衡熵和相关度的自动摘要技术研究
分类超曲面算法复杂度研究
Multi-view learning via probabilistic latent semantic analysis
An algorithm for solving rule sets-based bilevel decision problems
Cross-domain learning from multiple sources: A consensus regularization perspective
Exploiting associations between word clusters and document classes for cross-domain text categorizat
Adaptive associative classification with emerging frequent patterns
Evolutionary taxonomy construction from dynamic tag space
Recommending Flickr groups with social topic model
Parallel Web Mining System Based on Cloud Platform
基于摄动的模糊聚类算法最优模糊等价矩阵相关性质分析
PSO的向量整体修订策略和局部跳出策略
Triplex transfer learning: exploiting both shared and distinct concepts for text classification
CARSA: A context-aware reasoning-based service agent model for AI planning of web service compositio
Collaborative dual-PLSA: Mining distinction and commonality across multiple domains for text classif
An improved FCMBP fuzzy clustering method based on evolutionary programming
Learning semantic concepts from image database with hybrid generative/discriminative approach
一种基于跨领域典型相关性分析的迁移学习方法
基于语义轨迹的视频事件探测
基于维基百科和条件随机场的领域主题词抽取方法
基于动态描述逻辑的Web自动化服务组合
支持链式桥规则的分布式动态描述逻辑
基于内容图像检索中的优化鉴别特征
基于查询词出现的相关度改进
基于文档重要度的静态索引剪枝方法
Support vector machine with mixture of kernels for automatic image annotation
Fusing PLSA model and Markov random fields for automatic image annotation
基于用户浏览时间的点击模型
会议论文
Orthogonal nonnegative matrix tri-factorization for semi-supervised document co-clustering
DH-TRIE frequent pattern mining on Hadoop using JPA
Provenance-based Indexing Support in Micro-blog Platforms
Improving transfer learning by introspective reasoner
Parallel Max-Min Ant System Using MapReduce
D-LDA: A topic modeling approach without constraint generation for semi-defined classification
Support vector machine with mixture of kernels for image classification
Similarity-based Bayesian learning from semi-structured log files for fault diagnosis of Web service
Tweet Recommendation with Graph Co-Ranking
Parallel decision tree with application to water quality data analysis
Parallel CLARANS Clustering Based on MapReduce
SSHLDA: A Semi-Supervised Hierarchical Topic Model
Temporal and social context based burst detection from folksonomies
The Improved Non-negative Matrix Factorization Algorithm for Document Clustering
Parallel outlier detection using KD-tree based on MapReduce
Quad-tuple PLSA: Incorporating entity and its rating in aspect identification
The improved non-negative matrix factorization algorithm for document clustering
Detecting bursty events in collaborative tagging systems
Content-enriched classifier for web video classification
Evolutionary Taxonomy Construction from Dynamic Tag Space
Challenging the Long Tail Recommendation
Multi-Task Semi-Supervised Semantic Feature Learning for Classification
Parallel implementation of classification algorithms based on mapreduce
A parallel hyper-surface classifier for high dimensional data
Nonparametric curve extraction based on ant colony system
TINAC: A Fast and Effective Web Video Topic Detection Framework
Collaborative dual-PLSA: Mining distinction and commonality across multiple domains for text classif
Efficient Collecting Relevence Information from Clickthroughs for Web Retrieval System Evaluation
A Unified Model for Stable and Temporal Topic Detection from Social Media Data
On the support of ad-hoc semantic web data sharing
A noise handling method for hyper surface classification
Frequent term based peer-to-peer text clustering
Local Bayesian based rejection method for HSC ensemble
Embedding with autoencoder regularization
Exploiting associations between word clusters and document classes for cross-domain text categorizat
Parallel Max-Min Ant System Using MapReduce
Visualizing Timelines: Evolutionary Summarization via Iterative Reinforcement between Text and Image
Identifying Event-related Bursts via Social Media Activities
To Better Stand on the Shoulder of Giants: Learning to Identify Potentially Influential Literature
Citation Count Prediction: Learning to Estimate Future Citations for Literature
Timeline Generation through Evolutionary Trans-Temporal Summarization
Evolutionary Timeline Summarization: a Balanced Optimization Framework via Iterative Substitution
Topical Keyphrase Extraction From Twitter,
Comparing Twitter and Traditional Media using Topic Models
Effective Time Ratio: A Measure for Web Search Engines with Snippets
On modeling MapReduce with granular computing
Back propagation approach for semi-supervised learning in granular computing
Automatic image annotation with continuous PLSA
Mining hot clusters of similar anomalies for system management
Voting conditional random fields for multi-label image classification
Collaboration in agent grid based on dynamic description logics
Improving Large Graph Processing on Partitioned Graphs in the Cloud
Triplex transfer learning: Exploiting both shared and distinct concepts for text classification
Concept Learning for Cross-domain Text Classification: a General Probabilistic Framework
Combining Supervised and Unsupervised Models via Unconstrained Probabilistic Embedding
Outlier Detection Based on Extreme Learning Machine
A Visual Cognitive Method Based on Hyper Surface for Data Understanding
PPLSA: Parallel probabilistic latent semantic analysis based on MapReduce
Parallel implementation of apriori algorithm based on MapReduce
Effectively constructing reliable data for cross-domain text classification
The high-activity parallel implementation of data preprocessing based on mapreduce
A HSC-based sample selection method for Support Vector Machine
Effectively leveraging entropy and relevance for summarization
Context Modeling for Ranking and Tagging Bursty Features in Text Streams
PPS Sampling of Web Graph Using Preferential Jumping Strategy
HisTrace: A System for Mining on News-Related Articles instead of Web Pages
Mechanism of concept learning in topologized knowledge granular space based on GrC
Event recognition from news webpages through latent ingredients extraction
Parallel K-Means Clustering Based on MapReduce
Mining frequent agent action patterns for effective multi-agent-based web service composition
Multi-label image annotation based on neighbor pair correlation chain
Dynamic Reasoning under Probabilistic Uncertainty in the Semantic Web
Combining the missing link: An incremental topic model of document content and hyperlink
An efficient data indexing approach on Hadoop using Java persistence API
A mind model for brain-like computer
PAGE: A Partition Aware Graph Computation Engine
Shared structure learning for multiple tasks with multiple views
Multiple feature fusion for social media applications
Fast Time Series Classification Based on Infrequent Shapelets
Modeling User Expertise in Folksonomies by Fusing Multi-type Features
Keyword Query Reformulation on Structured Data
Summarize what you are interested in: An optimization framework for Interactive Personalized Summari
Finding a Wise Group of Experts in Social Networks
Jointly modeling aspects and opinions with a MaxEnt-LDA hybrid
TeRec: A Temporal Recommender System Over Tweet Stream
LCARS: A Location-Content-Aware Recommender System
A new similarity measure based on preference sequences for collaborative filtering
Optimized Top-K Processing with Global Page Scores on Block-Max Indexes,
专利
一种并行的垂直交叉网络数据采集方法及系统
著作
搜索引擎效果评测－基于用户点击日志分析的方法与技术
搜索引擎—原理、技术与系统（第二版）
相关项目
文本语义模型和子空间聚类研究
期刊论文 12 会议论文 11
面向个人数据同步和备份服务的数据中心存储节点优化技术研究
会议论文 3
WEB智能搜索中的文本内容信任判定方法研究
期刊论文 59 会议论文 11 专利 1
一种新型的Web知识获取方法及其应用研究
期刊论文 8
基于时空语义的Web信息检索研究
期刊论文 13 会议论文 14
基于儿童心理分析的图解式汉语口语自动解析方法研究
期刊论文 3
农业复杂自适应搜索模型研究
期刊论文 32 会议论文 3
李晓明的项目
基于网络社交媒体的层次化用户兴趣建模
期刊论文 8
基于IPv6的anycast路由的研究
期刊论文 2
第二届中美关系研讨会学术圆桌会议
第三届北京大学中美计算机科学高峰论坛
中美日“科学、社会与因特网”科技政策论坛
国家基础研究数据资源的开发及其在项目管理中的应用研究
MOOC数据模型及其对课程与学习评价的效用研究
期刊论文 3
互联网搜索和挖掘国际研讨会
海量Web数据中命名实体及相互关系提取与分析的模型和算法研究
期刊论文 8 会议论文 24 著作 2
SPMD程序设计模型——从Fortran到Java
Web结构和社会信息相关性模型及其计算实验方法研究
期刊论文 18 会议论文 6