"... In PAGE 12: ...-NN). In majority of cases, however, the performance of two classifiers dropped. As linear discriminant analysis performs poorly and k-NN only implicitly builds a classification model, we have further investigated only the classification trees ob- tained by C4.5 for the three best cases which are marked bold in Table1 . For those two that use the average activity of thigh muscles only (row 2) the decision tree con- structed using the whole data set was the same and consisted of one internal node only (single condition).... ..."

"... In PAGE 8: ...Table3... In PAGE 8: ...Table 3: Data statistics from 3am to 4am on November 3, 2005 The statistics in Table3 display the various source IP communicating with the FIU-SCIS network trying to communicate with the same destination IP on various distinct ports to find any vulnerable open ports on the destination IP targeted by the attacker. This graph displays the entire top five ... ..."

"... In PAGE 8: ...Table 2: Data statistics from 3 am to 4am on November 2nd, 2005 The statistics in Table2 display the various source IP communicating with the FIU-SCIS network on the destination port 22, trying to login into the system. Whenever the attacker is trying to login into the system using the brute force attack the amount of connections the particular Source IP exceeds the total amount of connection of any other IP by a huge margin.... ..."

"... In PAGE 3: .... where possible, interpret the resulting classi er. 20 di erent feature subsets were investigated. As for a relationship between the AS and the maneuver type, four possibilities were explored: sEMG obtained from hip/knee maneuver only, ankle maneuver only, hip/knee and ankle maneuvers combined, and an sEMG averaged across hip/knee and ankle maneuvers ( Table1 , column labels). Similarly, activity from ve recorded muscles was tested independently (q, a, h, ta, ts) or in four various combinations that include an average of only thigh muscles, an average of only leg muscles, an average of all muscles, or an average of thigh and leg muscles, respectively.... In PAGE 4: ...5. For the three best cases which are marked bold in Table1 , the decision trees are shown in Figure 1. For two best cases that use the average of the activity of thigh muscles (second row) the decision tree constructed using the whole data set was the same and consisted of one internal node (single condition) only (Figure 1.... ..."

"... In PAGE 6: ...RS Enforcement Period .............................................................. 86 Table33 .... In PAGE 99: ....5.1.2.3 Phase-In of A GRS (Preferred Alternative) Table33 shows the expected effects of Alternative 4 on the HT-CP sector in terms of retained harvest required to meet the GRS, the equivalent product weight, and additional product weight as a portion of total sector production.6 The analysis estimates that in 2005, only two vessels will need to increase its groundfish retention rate to meet the GRS for that year.... ..."

"... In PAGE 8: ...able 3-5: Summary of data for 2nd iteration.................................................................... 35 Table4 -1: Usability problems for reliability study.... In PAGE 8: ...able 4-1: Usability problems for reliability study........................................................... 43 Table4 -2: Example classification of Scenario#8 .... In PAGE 8: ...able 4-2: Example classification of Scenario#8 ............................................................. 49 Table4 -3: Results of user agreement at each level in the UAF.... In PAGE 8: ...able 4-3: Results of user agreement at each level in the UAF........................................ 50 Table4 -4: Results of overall user agreement .... In PAGE 8: ...able 4-4: Results of overall user agreement ................................................................... 51 Table4 -5: Strength of Agreement for various kappa values .... In PAGE 49: ..., 1999). Scenarios have been developed based on these usability problems and the participants were provided with the scenarios and instructions to classify (Appendix G) rather than a simple statement of a particular usability problem, as given in the Table4 -1. ... In PAGE 51: ...43 Table4 -1: Usability problems for reliability study Problem no. Type of usability problem Relevant area in User Action Framework 1 Unreadable error message Assessment 2 User does not understand master document feature Planning 3 User cannot find a feature to support re-using document numbers in a document retrieval system Translation 4 User clicks on wrong button Physical Action 5 User cannot directly change a file name in an FTP program Translation 6 User cannot tell if system is performing requested operation Assessment 7 User wants to fix database error but is confusing by button labels for appropriate action Translation 8 Program does not provide a Ctrl-P shortcut for printing Translation 9 User cannot understand the error message provided by system Assessment 10 Unnecessarily long error message Assessment 11 User cannot distinguish a button because of background Physical action 12 Data does not see way to select odd font size Translation 13 Data entry format not provided Translation 14 Uncontrollable scrolling Physical action 15 Vision impaired user needs preference options for setting larger font size Translation Procedure The reliability study was conducted at the Assessment and Cognitive Ergonomics Laboratory at Industrial and Systems Engineering department in Virginia Tech.... In PAGE 56: ... Reliability at each level within the hierarchical structure 2. Overall reliability for the terminal node Agreement at Different Levels in the UAF Table4 -2 shows an example of the data from one usability case description. Scenario 8 was about the lack of a specific short cut for a particular task.... In PAGE 56: ... Three participants (#2, #3 and #5) thought the scenario involved Existence. In order to continue measuring agreement accurately, the data from participants #2, #3 and #5 were eliminated (indicated by strikethrough effect in Table4 -2) from further reliability measures since these participants were now taking a different path than the remaining seven. At Level 3, all 7 of the remaining 7 participants agreed that the issue was about Preferences and Efficiency.... In PAGE 56: ... There were no more dis-agreements at other levels and all the remaining 7 participants agreed that the terminal node is about a specific short cut needed by the user. The example illustrated in Table4 -2 shows the approach for calculating reliability at different levels by eliminating participants that proceeded down a different path from the majority. This approach made sure that there were no continuous penalties for disagreement at lower levels when a participant was on a different path and had no opportunity to see the same choices as the other participants.... In PAGE 57: ...49 Table4 -2: Example classification of Scenario#8 Participant Level 1 Level 2 Level 3 Level 4 Level 5 1 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 2 Existence Preferences amp; Efficiency 3 Existence Existence of a way Existence of a feature 4 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 5 Existence Preferences amp; Efficiency 6 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 7 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 8 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 9 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 10 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted Results of reliability calculations across all scenarios for various levels are given in Table 4-3. Column 2 indicates the number of cases analyzed for each level within the UAF.... In PAGE 57: ...49 Table 4-2: Example classification of Scenario#8 Participant Level 1 Level 2 Level 3 Level 4 Level 5 1 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 2 Existence Preferences amp; Efficiency 3 Existence Existence of a way Existence of a feature 4 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 5 Existence Preferences amp; Efficiency 6 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 7 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 8 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 9 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted 10 Translation Task structure and interaction control Preferences amp; Efficiency Alternative ways to do tasks Specific short cut wanted Results of reliability calculations across all scenarios for various levels are given in Table4 -3. Column 2 indicates the number of cases analyzed for each level within the UAF.... In PAGE 57: ... For example, some scenarios required navigation down to only the third level in the UAF before terminal nodes were presented. As shown in Table4 -3, 6 cases required Level 5 classification while only 1 case required classification at 6th and 7th level. Values in the Po column indicate the proportion of observed agreement while values in the Pc column indicate the proportion of agreement ... In PAGE 58: ...50 decreases as the number of choices increase. As shown in Table4 -2, the proportion of chance agreement is higher at the top levels in the UAF than the lower levels because there are fewer choices at the top of the framework. Observed agreement requires substantially higher values to overcome chance agreement at the top levels of the UAF.... In PAGE 58: ... Observed agreement requires substantially higher values to overcome chance agreement at the top levels of the UAF. The kappa values shown in Table4 -3 (k column) indicated strong agreement at all levels within the UAF, especially at the top levels of the framework. The Z column contains the observed values for the standard normal variate obtained by dividing kappa by its standard error.... In PAGE 58: ... With such few cases, the agreement score would likely not be valid since the approximate normality assumption for kappa would be violated. Table4 -3: Results of user agreement at each level in the UAF Level Scenarios at this level Po Pc k Z*** 1 15 0.... In PAGE 58: ... The overall agreement provides reliability information for the various paths taken by each classifier. Kappa results for overall agreement ( Table4 -4) showed strong reliability (k = 0.610, p lt; .... In PAGE 59: ...051), considering the number of possible terminal nodes available to the users. Table4 -4: Results of overall user agreement Number of Scenarios Po Pc k Z*** 0.... In PAGE 60: ... The strength varies from Poor to Almost Perfect as kappa values vary from 0-1. Table4 -5 looks at various kappa values obtained in this study and categorizes them according to the scale provided by Landis and Koch (1977). These divisions provide some useful benchmarks to categorize kappa values, and interpret what it means in terms of users agreement.... ..."

"... In PAGE 7: ...able 3-9: GIOP and LW-IOP message structures..................................................................................... 87 Table4 -1: Comparison of Request pseudo interface between Orbix and Cool ORB.... In PAGE 7: ...able 4-1: Comparison of Request pseudo interface between Orbix and Cool ORB............................... 100 Table4 -2: Comparison of Server Request pseudo interface .... In PAGE 7: ...able 4-2: Comparison of Server Request pseudo interface .................................................................... 100 Table4 -3: Comparison of ORB pseudo interface .... In PAGE 7: ...able 4-3: Comparison of ORB pseudo interface .................................................................................... 101 Table4 -4: Comparison of BOA pseudo interface.... In PAGE 7: ...able 4-4: Comparison of BOA pseudo interface.................................................................................... 101 Table4 -5: Comparison of Object pseudo interface.... In PAGE 7: ...able 4-5: Comparison of Object pseudo interface.................................................................................. 102 Table4 -6: Comparison of NVList pseudo interface.... In PAGE 7: ...able 4-6: Comparison of NVList pseudo interface................................................................................. 102 Table4... In PAGE 100: ...one using Cool ORB R4.0 beta. and Orbix 2.0 [COOLa] and [ORBXb]. On the following tables the CORBA::Environment env default parameter is not shown. Request pseudo interface in C++ from CORBA Orbix Cool ORB Object_ptr target (); yes yes const char * operation (); yes yes NVList_ptr arguments (); yes yes NamedValue_ptr result (); yes yes Environment_ptr env(); yes yes ExceptionList_ptr exceptions (); no yes ContextList_ptr contexts (); no yes void ctx (Context_ptr); yes yes Context_ptr ctx (); yes yes Any amp;add_in_arg (); no yes Any amp;add_in_arg (const char* ); no yes Any amp;add_inout_arg (); no yes Any amp;add_inout_arg (const char*); no yes Any amp;add_out_arg (); no yes Any amp;add_out_arg (const char*); no yes void set_return_type (TypeCode_ptr ); no yes Any amp;return_value (); no yes Status invoke (); yes yes Status send_oneway (); yes yes Status send_deferred () yes no Status get_response (); yes no Boolean poll_response (); yes no Table4 -1: Comparison of Request pseudo interface between Orbix and Cool ORB ServerRequest pseudo interace in C++ by CORBA Orbix Cool ORB const char* op_name(); yes *) Operation_def_ptr op_def (); yes no Context_ptr ctx (); yes yes void params (NVList_ptr); yes yes Any* result (Any *value); no yes void exception (Any *value); yes yes Table 4-2: Comparison of Server Request pseudo interface... In PAGE 100: ...one using Cool ORB R4.0 beta. and Orbix 2.0 [COOLa] and [ORBXb]. On the following tables the CORBA::Environment env default parameter is not shown. Request pseudo interface in C++ from CORBA Orbix Cool ORB Object_ptr target (); yes yes const char * operation (); yes yes NVList_ptr arguments (); yes yes NamedValue_ptr result (); yes yes Environment_ptr env(); yes yes ExceptionList_ptr exceptions (); no yes ContextList_ptr contexts (); no yes void ctx (Context_ptr); yes yes Context_ptr ctx (); yes yes Any amp;add_in_arg (); no yes Any amp;add_in_arg (const char* ); no yes Any amp;add_inout_arg (); no yes Any amp;add_inout_arg (const char*); no yes Any amp;add_out_arg (); no yes Any amp;add_out_arg (const char*); no yes void set_return_type (TypeCode_ptr ); no yes Any amp;return_value (); no yes Status invoke (); yes yes Status send_oneway (); yes yes Status send_deferred () yes no Status get_response (); yes no Boolean poll_response (); yes no Table 4-1: Comparison of Request pseudo interface between Orbix and Cool ORB ServerRequest pseudo interace in C++ by CORBA Orbix Cool ORB const char* op_name(); yes *) Operation_def_ptr op_def (); yes no Context_ptr ctx (); yes yes void params (NVList_ptr); yes yes Any* result (Any *value); no yes void exception (Any *value); yes yes Table4 -2: Comparison of Server Request pseudo interface... In PAGE 101: ... All rights reserved Type: Internal Nature: Prototype Status: Intermediate Editing Status: Approved File: mpd3.doc ORB pseudo interface in C++ from CORBA Orbix Cool ORB string object_to_string (Object_ptr); yes yes Object string_to_object (const char*); yes yes Status create_list (Long, NVList_ptr amp;); yes yes Status create_operation_list (OperationDef_ptr, NVList_ptr amp;); yes no Status create_named_value (NamedValus_ptr amp;); no yes Status create_exception_list (ExceptionList_ptr amp;); no yes Status create_context_list (ContextList_ptr amp;); no yes Status get_default_context (Context_ptr amp;); yes yes Status create_environment (Environment_ptr amp;); yes yes Status send_multiple_requests_oneway (const RequestSeq amp;); yes no Status send_multiple_requests_defered (const RequestSeq amp;); yes no Boolean poll_next_response (); yes no Status get_next_response (Request_ptr amp;); yes no BOA BOA_init(); yes yes ObjectIdList list_initial_services (); yes, part of Naming Service package no, Object resolve_initial_references (const char*); yes, part of Naming Service package yes, very recently Table4 -3: Comparison of ORB pseudo interface BOA pseudo interface in C++ from CORBA Orbix Cool ORB Object create (const ReferenceData amp;, InterfaceDef_ptr, ImplementationDef_ptr); yes no void dispose (Object_ptr); yes yes ReferenceData* get_id (Object_ptr); yes no void change_implementation (Object_ptr, ImplementationDef_ptr); yes no Principal_ptr get_principal (Object_ptr, Environment_ptr); yes no void impl_is_ready (ImplementationDef_ptr) yes no void obj_is_ready (Object_ptr, ImplementationDef_ptr) yes no void deactivate_impl (ImplementationDef_ptr) yes no void deactivate_obj (Object_ptr) yes yes... In PAGE 101: ... All rights reserved Type: Internal Nature: Prototype Status: Intermediate Editing Status: Approved File: mpd3.doc ORB pseudo interface in C++ from CORBA Orbix Cool ORB string object_to_string (Object_ptr); yes yes Object string_to_object (const char*); yes yes Status create_list (Long, NVList_ptr amp;); yes yes Status create_operation_list (OperationDef_ptr, NVList_ptr amp;); yes no Status create_named_value (NamedValus_ptr amp;); no yes Status create_exception_list (ExceptionList_ptr amp;); no yes Status create_context_list (ContextList_ptr amp;); no yes Status get_default_context (Context_ptr amp;); yes yes Status create_environment (Environment_ptr amp;); yes yes Status send_multiple_requests_oneway (const RequestSeq amp;); yes no Status send_multiple_requests_defered (const RequestSeq amp;); yes no Boolean poll_next_response (); yes no Status get_next_response (Request_ptr amp;); yes no BOA BOA_init(); yes yes ObjectIdList list_initial_services (); yes, part of Naming Service package no, Object resolve_initial_references (const char*); yes, part of Naming Service package yes, very recently Table 4-3: Comparison of ORB pseudo interface BOA pseudo interface in C++ from CORBA Orbix Cool ORB Object create (const ReferenceData amp;, InterfaceDef_ptr, ImplementationDef_ptr); yes no void dispose (Object_ptr); yes yes ReferenceData* get_id (Object_ptr); yes no void change_implementation (Object_ptr, ImplementationDef_ptr); yes no Principal_ptr get_principal (Object_ptr, Environment_ptr); yes no void impl_is_ready (ImplementationDef_ptr) yes no void obj_is_ready (Object_ptr, ImplementationDef_ptr) yes no void deactivate_impl (ImplementationDef_ptr) yes no void deactivate_obj (Object_ptr) yes yes Table4... In PAGE 102: ...Table 4-5: Comparison of Object pseudo interface NVList pseudo interface in C++ from CORBA Orbix Cool ORB Ulong count (); yes yes NamedValue_ptr add (Flags); yes yes NamedValue_ptr add_item (const char*, Flags); yes yes NamedValue_ptr add_value (const char*, const Any amp;, Flags); yes yes NamedValue_ptr add_item_consume ( char*, Flags); no yes NamedValue_ptr add_value_consume (char*, Any*, Flags ); no yes NamedValue_ptr item (Ulong); yes yes Status remove (Ulong); yes yes Table4 -6: Comparison of NVList pseudo interface Context pseudo interface in C++ from CORBA Orbix Cool ORB const char* context_name() const yes yes Context_ptr parent () const yes yes Status create_child (const char*, Context_ptr amp;); yes yes Status set_one_value (const char*, const Any amp; ); yes yes Status set_values (NVList_ptr ); yes yes Status delete_values (const char* ); yes yes Status get_values (const char*, Flags, const char*, NVList_ptr amp; ); yes yes Table 4-7: Comparison of Context pseudo interface *) In Orbix a CORBA compliant feature is usually and explicitly written as such - it is not the case for... In PAGE 102: ...Table 4-5: Comparison of Object pseudo interface NVList pseudo interface in C++ from CORBA Orbix Cool ORB Ulong count (); yes yes NamedValue_ptr add (Flags); yes yes NamedValue_ptr add_item (const char*, Flags); yes yes NamedValue_ptr add_value (const char*, const Any amp;, Flags); yes yes NamedValue_ptr add_item_consume ( char*, Flags); no yes NamedValue_ptr add_value_consume (char*, Any*, Flags ); no yes NamedValue_ptr item (Ulong); yes yes Status remove (Ulong); yes yes Table 4-6: Comparison of NVList pseudo interface Context pseudo interface in C++ from CORBA Orbix Cool ORB const char* context_name() const yes yes Context_ptr parent () const yes yes Status create_child (const char*, Context_ptr amp;); yes yes Status set_one_value (const char*, const Any amp; ); yes yes Status set_values (NVList_ptr ); yes yes Status delete_values (const char* ); yes yes Status get_values (const char*, Flags, const char*, NVList_ptr amp; ); yes yes Table4 -7: Comparison of Context pseudo interface *) In Orbix a CORBA compliant feature is usually and explicitly written as such - it is not the case for... ..."

"... In PAGE 13: ... These country specific service links include both trade facilitation and the overall business climate. Table1 presents some service link and business climate indicators for selected developing countries.11 Although there is some variation within regions, the poor quality of service links in Africa, South Asia and some of Latin America emerge clearly.... In PAGE 17: ... Our key independent variable is an index of service link quality and costs. Service links were proxied by the synthetic measure of infrastructure consisting of transport, communications, and electric power reliability and costs, including those shown in Table1 , using data from around 2004. Appendix 1 provides a more complete description of the service link variable.... ..."