### Table 1. Robust estimation of the epipolar geometry from a set of matches containing outliers using RANSAC (a61 OK indicates the probability that the epipolar geometry has been correctly estimated).

"... In PAGE 4: ...a41a48a44 Our system incorporates the RANSAC (RANdom SAmpling Consesus)a96 approach used by Torr et al.a41a43a96 Table1 sketches this technique. Once the epipolar geometry has been retrieved, one can start looking for more matches to re ne this geometry.... In PAGE 5: ...his is illustrated in Fig. 3. The rst steps consists of nding the epipolar geometry as described in Section 3.1. Then the matches which correspond to already reconstructed points are used to compute the projection matrix a106 a100 . This is done using a robust procedure similar to the one laid out in Table1 . In this case a minimal sample of 6 matches is needed to compute a106 a100 .... ..."

### Table 1. Robust estimation of the epipolar geometry from a set of matches containing outliers using RANSAC (a61 OK indicates the probability that the epipolar geometry has been correctly estimated).

"... In PAGE 4: ...a41a48a44 Our system incorporates the RANSAC (RANdom SAmpling Consesus)a95 approach used by Torr et al.a41a43a95 Table1 sketches this technique. Once the epipolar geometry has been retrieved, one can start looking for more matches to refine this geometry.... In PAGE 5: ...his is illustrated in Fig. 3. The first steps consists of finding the epipolar geometry as described in Section 3.1. Then the matches which correspond to already reconstructed points are used to compute the projection matrix a104 a99 . This is done using a robust procedure similar to the one laid out in Table1 . In this case a minimal sample of 6 matches is needed to compute a104 a99 .... ..."

### Table 1. An example outlier detection test set (four major and one outlier topic)

"... In PAGE 6: ... The numbers of documents representing the outlier categories varied from 100% to 10% of the number of documents representing one major category in that test set. In Table1 we present an example outlier detection data set containing documents from one outlier category of size 30%. During the experiment, we fed all 77 data sets to the clustering algorithms and compared the contents of the automatically generated clusters with the reference categoriesdefinedinODP.... ..."

### Table 1 gives the descriptive statistics for the volatility of the o-c returns of the three indexes. Each of the volatility series exhibits clustering, which needs to be captured by an appropriate model. Furthermore, all series appear to contain a number of observations which might legitimately be regarded as outliers.

### Table 4. Directory-based outlier detection results.

"... In PAGE 9: ...2. Case Study Results The / and /usr/ partitions from the Forensic Chal- lenge were analyzed using the directory outlier analysis script and the results can be found in Table4 . There is only one directory in this system that contains hidden files and it is the /usr/man/.... ..."

### Table 3: The probability of being an outlier and the posterior mean of lo- cation shifts and standard deviations in the factor analysis (K = 2) with outliers model of the language data in Fuller (1987). Posterior means larger than posterior standard deviations are bold face.

1997

"... In PAGE 13: ... The Bayes factor is clearly in favor of the factor analysis model with outliers. Table3 is the summary of the important result from our factor analysis model with outliers. The rst column is the observation number and the second column is the probability of its being an outlier.... In PAGE 13: ... The third to the tenth columns contain the outlier shifts and the standard deviations of the outliers in parentheses. Table3 shows the row estimates of the location shift matrix A for which the posterior probability parameter quot;i (the probability of being an outlier) is larger than 1/2. The prior probability that observation i is an outlier is assumed to be quot;i = 0:1.... ..."

Cited by 1

### Table 3: Yates corrected 2{goodness{of{ t test with nine degrees of freedom for estimating the number of outliers. The rows show the values of the test statistic for the Data Sets A and Iris and clustering results with g clusters and n r outliers. See also the text.

2006

"... In PAGE 18: ... The in uence of the entropy term in the MAP{criterion (4) rather turns out to be too strong so that the estimates of the sizes break down. The 2{goodness{of{ t test presented on the right of Table3 indicates that Iris does not contain many outliers, if any.... ..."

### Table 1. Queries and the recall, precision, and effectiveness for each, given abstracts (Ab), sentences (Se), and phrases (Ph) as text units from which to extract interactions between the query terms or their synonyms, in MEDLINE abstracts containing both query terms. (The last query, an outlier, is discussed further in Appendix A.) Recall Precision Effectiveness Query terms

"... In PAGE 3: ...3 from MEDLINE using ten queries ( Table1 ) to its PUBMED interface.8 Each query was the AND of two biochemical nouns.... In PAGE 6: ... Information retrieval measures for different types of text units. Recall and precision figures are means over the relevant figures for each query (shown in Table1 for all text unit types except sentence pairs). Each figure was appropriately weighted, by the number of abstracts in the set associated with that query (in the case of precision of abstracts), the number of co-occurrences for that query within the text unit under consideration (in the case of precision of sentence pairs, sentences, and phrases), or by the number of interactions described for that query within the associated set of abstracts (for recall).... In PAGE 8: ... Effectiveness of sophisticated text processing techniques is higher than the baseline figures in Table 2 above for both the sentence and phrase text units. For phrases, sophisticated techniques led to an effectiveness higher than that of any entry in Table1 above. (However comparisons across reports should be interpreted with caution.... In PAGE 9: ...9 Appendix A: An Outlier Query It is interesting to consider an outlier from among our ten queries. For the query cholesterol AND flavonoid, smaller text units fared more poorly than for other queries ( Table1 ). Closer inspection of these abstracts showed that flavonoid is a large family of chemicals, and the name of a specific flavonoid is usually stated in the first sentence of an abstract.... ..."

### Table 1: Descriptive statistics before and after the removal of outliers Mean Median Minimum Maximum St. dev. Skewness Kurtosis

"... In PAGE 6: ... This amounts to 678 weekly returns, which are expressed as percentages. In Table1 we present some key statistics on these data. From the skewness and kurtosis it can easily be seen that these data (except perhaps for the DAX) contain outlying observations, as for example, the stock market crash of October 1987.... In PAGE 7: ...and NIKKEI and just one in case of the FTSE. After removal of these outliers, the resulting third and fourth moments improve, as can be seen from the second panel of Table1 . In our further analysis we will use the outlier corrected series.... ..."

### Table 8: Results of diagnosing the texture data by using a contaminated normal set. The number of outliers in the normal data set, and its corresponding percentage of the entire set are shown.

2006

"... In PAGE 7: ... As a result, the test set contains the rest of the examples from the chosen class (not in the normal set) plus 223 examples of non-outliers. Table8 shows the results of these experiments. The number of outliers used to contaminate the normal set is shown below the row with the values of K.... ..."

Cited by 3