Common use of Query Processing Clause in Contracts

Query Processing. ‌ The next set of computations is performed at query time. The first task is to identify the query-topic i.e. P (qj|ci) is computed as the ratio of number of occur- rences of qj in the topic file corresponding to cj to the total number of words in the file. If qj is absent in the topic file we use a smoothing value of one as the count of qj. Σ After computing the topic probabilities of the query, we compute the composite source rank (CSR) scores of sources based on the topical probabilities of the query. For a source sk, ▇▇▇▇ is given by the likelihood of the query belonging to representa- tive topic-classes. We treat this as a soft-classification problem. For a user query q and a set of representative = P (ci|q) × TSRki