The TREC-9 filtering track

Abstract.To meet the needs of education in the learning community, an adaptive filtering algorithm for teaching resources based on vector space model was proposed in the paper. First, feature selection and pseudo feedback were used to select the initial filtering profiles and thresholds through training algorithm. Then user feedback was utilized to modify the profiles and thresholds adaptively through filtering algorithm. The algorithm had two advantages, the first was that it could carry on self-study to improve the precision; the second was that the execution did not need massive initial texts in the process of filtering. The result manifested that the algorithm was effective.

show abstract

“…Then, we compute the logarithm mutual information between words and topics [4]. The formula 1 is shown as follows:…”

Section: Training Stagementioning

confidence: 99%

Research on Personalized Community E-Learning Recommendation Service System by Using Adaptive Filtering Algorithm

Liang

Luo

2007

Computational Science – ICCS 2007

View full text Add to dashboard Cite

show abstract

“…Since then there has been a rapid transition to statistical models among researchers studying this area. The best forum for research on the topic has been the TREC Filtering tracks during TREC-8 (Hull and Robertson, 1999), TREC-9 (Robertson and Hull, 2001), and TREC-10 (Robertson and Soboroff, 2002).…”

Section: Adaptive Information Filteringmentioning

confidence: 99%

“…2 This process continued until each document in the stream had been considered. The system's performance over the entire stream of documents was measured using a variety of measures, including Precision, Recall, and various Utility metrics (Hull and Robertson, 1999;Robertson and Hull, 2001). This filtering environment is not a perfect model of an operational environment, but it is perhaps the best that can be done in the absence of real data and real users.…”

Section: Adaptive Information Filteringmentioning

confidence: 99%

“…The magnitude of an increase to a threshold is limited by maxstep , which is empirically set to 0.005, so that a non-relevant document with an extremely high score will not push the threshold too high, thus controlling the effects of outliers. Profile performance was measured with the TREC-8 F2 utility metric (Hull and Robertson, 1999). If a profile's F2 utility is positive, it is regarded as a good profile, and, therefore its threshold is decreased comparatively faster ( Figure 3.4).…”

Section: Up-and-downmentioning

confidence: 99%

“…It was previously used by the TREC8 Filtering Track (Hull and Robertson, 1999). TREC topics 351-400 were used to simulate user profiles.…”

Section: Datamentioning

confidence: 99%

See 2 more Smart Citations

Getting What You Want: Accurate Document Filtering in a Terabyte World

Callan¹

2002

View full text Add to dashboard Cite

REPORT DOCUMENTATION PAGE Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing this collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden to

show abstract

Modeling user interest shift using a bayesian approach

Lam¹,

Mostafa²

2001

J. Am. Soc. Inf. Sci.

View full text Add to dashboard Cite

We investigate the modeling of changes in user interest in information filtering systems. A new technique for tracking user interest shifts based on a Bayesian approach is developed. The interest tracker is integrated into a profile learning module of a filtering system. We present an analytical study to establish the rate of convergence for the profile learning with and without the user interest tracking component. We examine the relationship among degree of shift, cost of detection error, and time needed for detection. To study the effect of different patterns of interest shift on system performance we also conducted several filtering experiments. Generally, the findings show that the Bayesian approach is a feasible and effective technique for modeling user interest shift.

show abstract

The TREC-9 filtering track

Cited by 19 publications

References 0 publications

Research on Personalized Community E-Learning Recommendation Service System by Using Adaptive Filtering Algorithm

Research on Personalized Community E-Learning Recommendation Service System by Using Adaptive Filtering Algorithm

Getting What You Want: Accurate Document Filtering in a Terabyte World

Modeling user interest shift using a bayesian approach

Contact Info

Product

Resources

About