| A00-1004 | is still an experimental area . | Measuring | the confidence values of an alignment |
| A88-1020 | synonyms may prove to be useful is in | measuring | the semantic distance between |
| A83-1031 | discussed below in the section | Measuring | System Per - formance . Some |
| A00-1020 | of SWIZZLE can be achieved by | measuring | its impact on several complex |
| A92-1025 | Second , it provides an easy way of | measuring | success , by comparing system |
| A83-1028 | convenes . C. Qualitative Performance | Measuring | MT system throughput is one thing |
| A97-1050 | adopt the following approach to | measuring | recall : the upper bound is defined |
| A97-1056 | strategies and evaluation criteria for | measuring | model fit . We restrict the selection |
| A00-2038 | Once an appropriate function for | measuring | similarity between pairs of segments |
| A00-1004 | Therefore , caution is required when | measuring | structure difference numerically |
| A00-2038 | main components : a metric for | measuring | distance between phones , and |
| A97-1042 | sentence position in the corpus , | measuring | against the topic keywords . |
| A92-1035 | unit facet selects an appropriate | measuring | unit from the class TEMPORAL-UNIT |
| A00-1045 | we fruitfully apply the idea of | measuring | the coverage of a set of test |
| A97-1056 | training sample is determined by | measuring | the fit of the model to the sample |
| A92-1035 | constraint is not violated . The | measuring | - unit facet selects an appropriate |
| A88-1007 | to succeed by default . 5.2 . | Measuring | the System 's Learning As more |
| A97-1050 | emphasize quite clearly that K is | measuring | not the level of absolute agreement |
| A88-1020 | aspect and that an algorithm for | measuring | semantic distance between words |
| A88-1020 | senses and an automatic means for | measuring | semantic distance . Properties |
