Probabilistic question answering on the web

Radev, Dragomir; Fan, Weiguo; Qi, Hong; Wu, Hao; Grewal, A.

doi:10.1145/511446.511500

Cited by 90 publications

(77 citation statements)

References 12 publications

(2 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We achieved mean reciprocal rank of the answer (MRR) of 0.314, which is smaller than reported in [2], but seemingly better (although not directly comparable due to different testing methodologies) to the results reported with the other approaches [3][1] that were completely trainable (not relying on hand-crafted rules). We believe that the lower results are due to our use of AltaVista instead of Google and due to the fact that we did not use any manually crafted semantic filters.…”

Section: Empirical Evaluation and Conclusioncontrasting

confidence: 51%

Self-learning web question answering system

Roussinov

Robles

2004

Alternate Track Papers &Amp; Posters of the 13th International Conference on World Wide Web - WWW Alt. '04

View full text Add to dashboard Cite

While being quite successful in providing keyword based access to web pages, commercial search portals, such as Google, Yahoo, AltaVista, and AOL, still lack the ability to answer questions expressed in a natural language. In this paper, we present a probabilistic approach to automated question answering on the Web. Our approach is based on pattern matching and answer triangulation. By taking advantage of the redundancy inherent in the Web, each answer found by the system is triangulated (confirmed or disconfirmed) against other possible answers. Our approach is entirely self-learning: it does not involve any linguistic resources, nor it does require any manual tuning. Thus, the propose approach can easily be replicated in other information systems with large redundancy.

show abstract

Section: Empirical Evaluation and Conclusioncontrasting

confidence: 51%

Self-learning web question answering system

Roussinov

Robles

2004

Alternate Track Papers &Amp; Posters of the 13th International Conference on World Wide Web - WWW Alt. '04

View full text Add to dashboard Cite

show abstract

“…In fact the main objective of the effort was how the system carried out its ranking processes according to a set of techniques. These techniques included the proximity of the algorithm use and the probabilistic phrase ranking technique used by the system [21].…”

Section: Literature Review Andmentioning

confidence: 99%

Assessment of Measures for Information Retrieval System Evaluation: A Usercentered Approach

Akhigbe¹,

Afolabi²,

Adagunodo³

2011

IJCA

View full text Add to dashboard Cite

The ever increasing need for Information globally is a primary reason for the scores of daily usage of IR system. Therefore there is a need to evaluate the system from a more holistic perspective -of both the system and the user. At the moment system-centered measures are not usable for the user-centered approach. Therefore, this paper attempts to determine and also suggest measures as well as methods to meet this need. The factor analytic technique was experimented for this purpose, and the structural equation modeling technique was used to estimate the resultant model. Results show that the study demonstrated high significance. Hence, the statistics presented is capable of inspiring further work in IR systems' evaluation from user's perspective.

show abstract

“…Rules created for a specific question ontology must be re-tailored before being applied to different ontologies. In addition to TREC QA track systems, several web-based QA systems have relied on such rules with limited success [26]. Therefore, there is a need for more robust systems that can easily be adapted to handle new data sets and question ontologies.…”

Section: Related Workmentioning

confidence: 99%

Analysis of Statistical Question Classification for Fact-Based Questions

2005

View full text Add to dashboard Cite

Abstract. Question classification systems play an important role in question answering systems and can be used in a wide range of other domains. The goal of question classification is to accurately assign labels to questions based on expected answer type. Most approaches in the past have relied on matching questions against hand-crafted rules. However, rules require laborious effort to create and often suffer from being too specific. Statistical question classification methods overcome these issues by employing machine learning techniques. We empirically show that a statistical approach is robust and achieves good performance on three diverse data sets with little or no hand tuning. Furthermore, we examine the role different syntactic and semantic features have on performance. We find that semantic features tend to increase performance more than purely syntactic features. Finally, we analyze common causes of misclassification error and provide insight into ways they may be overcome.

show abstract

Probabilistic question answering on the web

Cited by 90 publications

References 12 publications

Self-learning web question answering system

Self-learning web question answering system

Assessment of Measures for Information Retrieval System Evaluation: A Usercentered Approach

Analysis of Statistical Question Classification for Fact-Based Questions

Contact Info

Product

Resources

About