"... In PAGE 3: ... Moreover CAS ex- plains the usages of KINVS. Table3 shows the listings of explanation. Figure 4 shows the example of the explana- tion of CAS, which is an explanation of WR.... ..."

"... In PAGE 19: ... These options will be made clearer through examples presented in the following sections. Situated explanation is summarized in Table3 . Unlike some knowledge acquisition ap- proaches, it does not include an explicit check for consistency when newly learned knowledge is added to the agent apos;s knowledge base.... ..."

"... In PAGE 6: ... Using the odds ratio we can then determine the efiect of the change. The kind of dialog that can be produced can be seen in Table2 . In this sample explanation we can see the advantage of using our local model to classify the Query Case as this gives us a measure of confldence in the prediction.... ..."

"... In PAGE 12: ...o binary causal models (i.e., each endogenous variable may take only two values). Our complexity results on these two problems for the general and the binary case are summarized in Table1 . In detail, the problem Explanation is complete for the class a26a17a27 a5 (resp.... In PAGE 12: ... On the other hand, recognizing explanations is only mildly harder than recognizing weak causes, which is a28 a29 a5 -complete. We now show how the complexity results in Table1 can be formally derived. In order not to distract from the flow of reading, we present the main parts and key ideas behind constructions, and move some technical details to Appendix A.... ..."

"... In PAGE 11: ... The research demonstrates that IS research articles have a variety of distinct investigative purposes and produce different types of explanation. There were three research types represented in this sample ( Table4 ); 49 articles present either designs for systems, frameworks for research or mathematical models (30% of total articles). These research articles were not intended to produce explanatory results of the types distinguished by the philosophy of science but rather present artifact designs, research programs or representations (models) of phenomena.... ..."

"... In PAGE 4: ... Table2 . Soil attributes used in analysis, their abbreviations, and the range of values found across the State.... In PAGE 4: ... Nonetheless, my purpose here was to un- derstand general landscape relationships over a large area and not to convey information about in- dividual small parcels of land. Landscape attributes used for comparisons with land types included percent sand and clay at two depths, available water to 152 cm (60 inches), per- cent organic matter, permeability, drainage class, thickness of soil solum, productivity index, and maximum slope ( Table2 ). Productivity index values integrate several soil characteristics and can be roughly equated to yield (bushels of corn per acre (Ode11 and Oschwald 1970)).... In PAGE 12: ...uble5. Regression results comparing the percentage of each land type or change in each soil association to landscape attributes. Based on 50 soil associations. See Table2 for explanations of variables. Class Average percent RZ per association A.... ..."

"... In PAGE 3: ... E measures the agreement between and yi. Table1 gives a summary of notations described above. Next, as examples of performance measures, we present the def- initions of MAP and NDCG.... ..."

"... In PAGE 3: ... E measures the agreement between and yi. Table1 gives a summary of notations described above. Next, as examples of performance measures, we present the def- initions of MAP and NDCG.... ..."