### Table 3: Discretized data set.

2007

"... In PAGE 5: ...0 j y1;1 y1;2 y1;3 y2;1 y2;2 y3;1 y3;2 y3;3 z Binary Variables 1 1 1 1 1 1 0 0 1 1 2 0 1 1 0 0 1 1 1 1 3 0 0 0 0 1 0 1 1 1 4 1 1 1 0 0 0 1 1 0 5 0 0 1 0 1 0 0 0 0 A discretized data set is then created; it has the same dimension as the original experiment set and whose components xj i, thereafter called variables, are mapped to the observables. The discretization of the binarized data set is carried out as follows: xj i = K(i) X k=1 yj i;k; for i = 1; : : : ; s; j = 1; : : : ; N: Table3 below displays the discretized data set. Positive (resp.... ..."

### Table 8 : Discretized Data Set

"... In PAGE 13: ... The data set contains supplier capability and supplier performance information on 23 suppliers. Ten attributes (shown in Table8 ) are identified (Talluri and Narasimhan, 2004) including quality management practices and systems (QMP), documentation and self-audit (SA), process/manufacturing capability (PMC), management of firm (MGT), design and development capabilities (DD), cost (C), quality (Q), price (P), delivery (D), cost reduction performance (CRP) and others. Talluri and Narasimhan (2004) apply Data Envelopment Analysis (DEA) to determine the efficiency of each supplier.... In PAGE 13: ... Talluri and Narasimhan (2004) apply Data Envelopment Analysis (DEA) to determine the efficiency of each supplier. Their conclusion on each supplier is shown in the last column of Table8 . All suppliers with efficiency (shown as Effi column) equal to one are considered ... ..."

### Table 1. Discrete data X continuous media

### Table 1. Clustering accuracy summary for paired samples. The di erence series is the continuous data minus the discrete data

2003

"... In PAGE 10: ... With the paired tests, the k-means algorithm is run (u = 20 times) on continuous series and discrete series derived from the same continuous data. Table1 shows that there is little di erence in the accuracy of the resulting clusterings. Table 1.... ..."

Cited by 7

### Table 1: Experimental results from the original and from the discretized data with features f=12

### Table 8. The training times for discretized data sets of different classification methods. The unit is second.

"... In PAGE 9: ... Since all compared algorithms are im- plemented with the Java language and all experiments are performed on the same computer, the comparisons of their efficiency are meaningful. From Table8 , it can be seen that the DFL algorithm is more efficient than other methods in most cases. 6 Discussion The fundamental difference between the DFL algorithm and other classification methods lies in the underlying phi- losophy of the algorithms, as shown in Figure 6.... ..."

### Table 2 Classical rough set-based feature selection where numerical attributes are discretized Data Feature CART

2006

"... In PAGE 7: ...introduced to evaluate the quality of selected features. Table2 shows the results with classical rough method [36], where N1 and N2 mean number of features in original data and reduced data, respectively. Table 3 gives the results based on fuzzy information entropy reduction.... ..."

### Table 9: Results of the application of MMR on discretized Data Set Cluster Number Object Number Efficiency

### Table 2: Subgroup descriptions induced by algorithm CN2- SD on discretized data. The algorithm induced 20 descriptions, but only three of those are different.

### Table 1. Results obtained on gaussian and discrete data. The accuracy confidence intervals was computed by repeating the data generation, learning and classification steps 100 times for gaussian variables and 60 times for discrete variables.

"... In PAGE 10: ... 5.2 Results The results obtained with the Gaussian model and for the discrete model are shown in Table1 . Accuracy and Brier score are reported with their 95% confi- dence intervals.... ..."