A General Evaluation Framework for Topical Crawlers

Topical crawlers are becoming important tools to support applications such as specialized Web portals, online searching, and competitive intelligence. As the Web mining field matures, the disparate crawling strategies proposed in the literature will have to be evaluated and compared on common tasks through well-defined performance measures. This paper presents a general framework to evaluate topical crawlers. We identify a class of tasks that model crawling applications of different nature and difficulty. We then introduce a set of performance measures for fair comparative evaluations of crawlers along several dimensions including generalized notions of precision, recall, and efficiency that are appropriate and practical for the Web. The framework relies on independent relevance judgements compiled by human editors and available from public directories. Two sources of evidence are proposed to assess crawled pages, capturing different relevance criteria. Finally we introduce a set of topic characterizations to analyze the variability in crawling effectiveness across topics. The proposed evaluation framework synthesizes a number of methodologies in the topical crawlers literature and many lessons learned from several studies conducted by our group. The general framework is described in detail and then illustrated in practice by a case study that evaluates four public crawling algorithms. We found that the proposed framework is effective at evaluating, comparing, differentiating and interpreting the performance of the four crawlers. For example, we found the IS crawler to be most sensitive to the popularity of topics.Partially supported by National Science Foundation CAREER grant No. IIS-0133124/0348940.     

Long-term trends in home advantage in professional team sports in North America and England (1876-2003)

Home advantage is quantitatively defined and calculated for each season since the start of the main professional sports in North America and England. Over 400,000 games are analysed. The leagues represented are the National League (1876-2002) and American League (1901-2002) for baseball, the National Hockey League (1917-2003) for ice hockey, the National Football League (1933-2002) for American football, the National Basketball Association (1946-2003) for basketball, and the four levels of professional football, formerly called the Football League, in England (1888-2003). Problems caused by unbalanced playing schedules are considered. The results are presented graphically to show long-term trends and sudden changes. The highest levels of home advantage for all sports were in their early years of existence. Home advantage in ice hockey, basketball and football in England has declined over the last two decades. In baseball there has been very little change over the last 100 years, with home advantage consistently lower than in other sports. There was a large drop in home advantage in football in England following the 7-year suspension of the league during the Second World War. The trends and changes provide some evidence that travel and familiarity contribute to home advantage, but little in support of crowd effects.     

Relationship between laboratory-measured variables and heart rate during an ultra-endurance triathlon

The aim of the present study was to examine the relationship between the performance heart rate during an ultra-endurance triathlon and the heart rate corresponding to several demarcation points measured during laboratory-based progressive cycle ergometry and treadmill running. Less than one month before an ultra-endurance triathlon, 21 well-trained ultra-endurance triathletes (mean +/- s: age 35 +/- 6 years, height 1.77 +/- 0.05 m, mass 74.0 +/- 6.9 kg, = 4.75 +/- 0.42 l x min(-1)) performed progressive exercise tests of cycle ergometry and treadmill running for the determination of peak oxygen uptake (VO2peak), heart rate corresponding to the first and second ventilatory thresholds, as well as the heart rate deflection point. Portable telemetry units recorded heart rate at 60 s increments throughout the ultra-endurance triathlon. Heart rate during the cycle and run phases of the ultra-endurance triathlon (148 +/- 9 and 143 +/- 13 beats x min(-1) respectively) were significantly (P < 0.05) less than the second ventilatory thresholds (160 +/- 13 and 165 +/- 14 beats x min(-1) respectively) and heart rate deflection points (170 +/- 13 and 179 +/- 9 beats x min(-1) respectively). However, mean heart rate during the cycle and run phases of the ultra-endurance triathlon were significantly related to (r = 0.76 and 0.66; P < 0.01), and not significantly different from, the first ventilatory thresholds (146 +/- 12 and 148 +/- 15 beats x min(-1) respectively). Furthermore, the difference between heart rate during the cycle phase of the ultra-endurance triathlon and heart rate at the first ventilatory threshold was related to marathon run time (r = 0.61; P < 0.01) and overall ultra-endurance triathlon time (r = 0.45; P < 0.05). The results suggest that triathletes perform the cycle and run phases of the ultra-endurance triathlon at an exercise intensity near their first ventilatory threshold.     

Stability of pedalling mechanics during a prolonged cycling exercise performed at different cadences

The aim of this study was to analyse the effect of pedalling rate on the pattern of mechanical torque application and on neuromuscular fatigue during prolonged cycling exercise. Eleven well-trained individuals performed three 1-h pedalling sessions, at 50 rev.min-1, 110 rev.min-1 and a freely chosen cadence, at an intensity corresponding to 65% of their maximal aerobic power. The mechanical torque applied on the right pedal was recorded for 30 s every 5 min while pedalling. Contractile and neural properties of the quadriceps and hamstring muscles were analysed before and immediately after each of the three pedalling sessions. The post-exercise reduction in knee extensors maximal voluntary contraction was significant (P<0.01) irrespective of the cadence, but no difference was found between cadences. The use of a particular cadence did not lead to preferentially central or peripheral fatigue. An increase in cadence resulted in greater positive and negative work generated during pedalling. The mechanical pattern was not altered during the exercise, whatever the selected cadence. The present study demonstrates that despite the occurrence of neuromuscular fatigue, trained individuals maintained a stable pedalling pattern throughout an endurance cycling exercise for cadences ranging from 50 to 110 rev.min-1.     

Competitive State Anxiety Inventory-2 (CSAI-2): evaluating the Swedish version by confirmatory factor analyses

The Competitive State Anxiety Inventory-2 (CSAI-2) is one of the most frequently used instruments when assessing competitive state anxiety in sport psychology research. However, doubts have been expressed about the factorial validity of both the English and the Greek versions of the scale. Hence, a revised version of the inventory (CSAI-2R) has recently been suggested to be more psychometrically sound (Cox et al., 2003). In the present study, the aim was to evaluate the psychometric properties of the Swedish version of the CSAI-2 using confirmatory factor analyses. A total of 969 athletes (571 men and 398 women) competing in 26 different sports completed the Swedish version of the CSAI-2. Three different factor structures were evaluated: the original three-factor model (with cognitive anxiety, somatic anxiety and self-confidence), a two-factor model in which self-confidence was excluded, and a three-factor model containing 17 items (CSAI-2R). The results revealed that only the 17-item model displayed an acceptable fit to the data. Although some doubts remain about the amount of variance that can be attributed to error variance in the subscales, the results suggest that it is better to use the CSAI-2R rather than the original CSAI-2.     

Visual principles for training by television

Educational technology research and development -