SVM based adaptive learning method for text classification from positive and unlabeled documents

Peng, Tao; Zuo, Wangmeng; He, Fei

doi:10.1007/s10115-007-0107-1

Cited by 62 publications

(29 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…7. We observed that the F-Measures of our proposed incremental SVM method outperform PSOC [29] and SVM which are two typical classi¯cation algorithms. Besides, without clustering, only using classi¯er to guide the topical crawling could get the results in shorter time.…”

Section: The Comparison Of Di®erent Techniquesmentioning

confidence: 73%

“…This incremental SVM modi¯es SVM [8] formulation so that the¯nal classi¯er has higher precision. In this paper, our previous work 1-DNFII algorithm [29], which considers both the diversity of the feature frequency in positive feature set and unlabeled example set and the absolute frequency of the feature in positive feature set, is used for collecting the positive and negative feature sets. If the frequency of occurrence of a feature in the positive example set is more frequently than in the unlabeled example set, then this feature is regarded as a positive example.…”

Section: Building Incremental Classi¯er For Topical Web Crawlingmentioning

confidence: 99%

See 1 more Smart Citation

Clustering-Based Topical Web Crawling for Topic-Specific Information Retrieval Guided by Incremental Classifier

Peng

Liu

2015

Int. J. Soft. Eng. Knowl. Eng.

Self Cite

View full text Add to dashboard Cite

Today more and more information on the Web makes it di±cult to get domain-speci¯c information due to the huge amount of data sources and the keywords that have few features. Anchor texts, which contain a few features of a speci¯c topic, play an important role in domain-speci¯c information retrieval, especially in Web page classi¯cation. However, the features contained in anchor texts are not informative enough. This paper presents a novel incremental method for Web page classi¯cation enhanced by link-contexts and clustering. Directly applying the vector of anchor text to a classi¯er might not get a good result because of the limited amount of features. Link-context is used¯rst to obtain the contextual information of the anchor text. Then, a hierarchical clustering method is introduced to cluster feature vectors and content unit, which increases the length of a feature vector belonging to one speci¯c class. Finally, incremental SVM is proposed to get the¯nal classi¯er and increase the accuracy and e±ciency of a classi¯er. Experimental results show that the performance of our proposed method outperforms the conventional topical Web crawler in Harvest rate and Target recall.

show abstract

Section: The Comparison Of Di®erent Techniquesmentioning

confidence: 73%

Section: Building Incremental Classi¯er For Topical Web Crawlingmentioning

confidence: 99%

Clustering-Based Topical Web Crawling for Topic-Specific Information Retrieval Guided by Incremental Classifier

Peng

Liu

2015

Int. J. Soft. Eng. Knowl. Eng.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Afterwards, we used k-means++ to classify all the activities into 10 clusters. Then we adopted the F 1 measure for performance evaluation [24], which is commonly used in text classification. The best topic number Z should correspond to the highest F 1 value.…”

Section: Parameter Learningmentioning

confidence: 99%

Predicting activity attendance in event-based social networks

Mei

et al. 2014

Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing

105

View full text Add to dashboard Cite

The newly emerging event-based social networks (EBSNs) connect online and offline social interactions, offering a great opportunity to understand behaviors in the cyber-physical space. While existing efforts have mainly focused on investigating user behaviors in traditional social network services (SNS), this paper aims to exploit individual behaviors in EBSNs, which remains an unsolved problem. In particular, our method predicts activity attendance by discovering a set of factors that connect the physical and cyber spaces and influence individual's attendance of activities in EBSNs. These factors, including content preference, context (spatial and temporal) and social influence, are extracted using different models and techniques. We further propose a novel Singular Value Decomposition with Multi-Factor Neighborhood (SVD-MFN) algorithm to predict activity attendance by integrating the discovered heterogeneous factors into a single framework, in which these factors are fused through a neighborhood set. Experiments based on real-world data from Douban Events demonstrate that the proposed SVD-MFN algorithm outperforms the state-of-the-art prediction methods.

show abstract

“…Aunque el algoritmo resultante es formalmente similar, cada producto vectorial, resultante de la representación cuadrática se sustituye por una función de núcleo nolineal. Los SVM se utilizan ampliamente y han conseguido lograr muy buenos resultados en categorización de contenido textual [29] .…”

Section: Máquinas De Soporte Vectorialunclassified

Modelo predictivo de control en fundiciones de alta precisión: un nuevo enfoque para la fase de predicción

Nieves¹,

Santos²,

Bringas³

2011

REVMETAL

View full text Add to dashboard Cite

Se tiene conocimiento que desde épocas ancestrales existen los procesos productivos. Procesos que han quedado representados tanto en versos bíblicos y pinturas egipcias, como en los grabados de las vasijas griegas. A lo largo de todo este tiempo, los procesos productivos han permitido transformar las materias primas en productos elaborados. Antiguamente, el proceso era mucho más primitivo, sin embargo, en la actualidad, y debido al avance y evolución de la sociedad, el proceso se ha mejorado. Los cambios más relevantes que se han producido a lo largo de las últi-mas décadas son: (i) la competencia ha aumentado drásticamente y (ii) las limitaciones impuestas por la legislación medioambiental han dotado de una complejidad extra al proceso de producción.Estos cambios implican que se lleve a cabo un control estricto de los procesos de producción. De esta forma, y gracias a la ayuda de múltiples disciplinas del conocimiento, es posible anticiparse y solucionar las situaciones adversas del proceso productivo. Así, los aspectos fundamentales sobre los que se puede trabajar son los siguientes: (i) minimizar los costes, (ii) maximizar la tasa de producción, (iii) Palabras claveModelo predictivo de control; Aprendizaje automático; Predicción de defectos; Minería de datos; Optimización de procesos. Model predictive control on high precision foundries: a new approach for the prediction phase AbstractA Model Predictive Control (MPC) is a system which allows us to control a production plant. Thanks to this type of system is possible to make a production that comes close to "zero defects". In order to achieve its main goal, this kind of systems consists of several phases. One of the most important is the phase that predicts the situation in which the plant is going to be in a given time. Currently, the majority of the research in this field are related to linear MPC, although the process, which the model tries to represent, may not be. Thus, this paper presents several experiments that proof that the forecast phase, usually represented by a single mathematical function, can be represented by machine-learning models. KeywordsModel predictive control; Machine learning; Fault prediction; Data mining; Process optimisation. minimizar el stock, (iv) controlar la fluctuación de los precios y (v) evitar fallos en el servicio del producto debidos a defectos ocultos. Esta situación también está afectando al proceso de producción de piezas de metal o proceso metalúrgico. Esta industria es una parte importante de la evolución de nuestra sociedad. Ya desde el punto de vista histórico se ha relacionado el proceso de producción metalúrgico con el desarrollo de la civilización humana, considerándolo incluso como una de las profesiones más antiguas de la historia. La fundición provee a otras industrias, como a la industria automovilística o naval, entre otras. Debido a la naturaleza de estas industrias, hay que realizar grandes controles de seguridad que aseguren el funcionamiento final de las piezas producidas, ya que el menor de los defecto...

show abstract

SVM based adaptive learning method for text classification from positive and unlabeled documents

Cited by 62 publications

References 14 publications

Clustering-Based Topical Web Crawling for Topic-Specific Information Retrieval Guided by Incremental Classifier

Clustering-Based Topical Web Crawling for Topic-Specific Information Retrieval Guided by Incremental Classifier

Predicting activity attendance in event-based social networks

Modelo predictivo de control en fundiciones de alta precisión: un nuevo enfoque para la fase de predicción

Contact Info

Product

Resources

About