Representing Queries as Distributions
نویسندگان
چکیده
Representing a query appropriately helps model the underlying information need and thus improves the retrieval performance. Previous query representations either generate related words and phrases to augment the original query but ignore how these words and phrases fit together in new queries, or apply a specific reformulation operation to the original query but ignore alternative operations. In this paper, a novel representation is proposed as a distribution of queries, where each query is a variation of the original query. This representation, on one hand, considers a query as a basic unit and thus captures important dependencies between words and phrases in the query. On the other hand, it naturally combines different reformulation operations as possible ways to generate variations of the original query. This query distribution representation is carefully compared with previous query representations in this paper to show its advantages. Some recent work using this representation has shown promising results and is briefly described here.
منابع مشابه
Learning Markov Network Structures Constrained by Context-Specific Independences
This work focuses on learning the structure of Markov networks. Markov networks are parametric models for compactly representing complex probability distributions. These models are composed by: a structure and numerical weights. The structure describes independences that hold in the distribution. Depending on the goal of learning intended by the user, structure learning algorithms can be divide...
متن کاملProbabilistic Reuse of Past Search Results
In this paper, a new Monte Carlo algorithm to improve precision of information retrieval by using past search results is presented. Experiments were carried out to compare the proposed algorithm with traditional retrieval on a simulated dataset. In this dataset, documents, queries, and judgments of users were simulated. Exponential and Zipf distributions were used to build document collections....
متن کاملEfficient Probabilistic Inference with Partial Ranking Queries
Distributions over rankings are used to model data in various settings such as preference analysis and political elections. The factorial size of the space of rankings, however, typically forces one to make structural assumptions, such as smoothness, sparsity, or probabilistic independence about these underlying distributions. We approach the modeling problem from the computational principle th...
متن کاملE cient Probabilistic Inference with Partial Ranking Queries
Distributions over rankings are used to model data in various settings such as preference analysis and political elections. The factorial size of the space of rankings, however, typically forces one to make structural assumptions, such as smoothness, sparsity, or probabilistic independence about these underlying distributions. We approach the modeling problem from the computational principle th...
متن کاملImprovement of the Analytical Queries Response Time in Real-Time Data Warehouse using Materialized Views Concatenation
A real-time data warehouse is a collection of recent and hierarchical data that is used for managers’ decision-making by creating online analytical queries. The volume of data collected from data sources and entered into the real-time data warehouse is constantly increasing. Moreover, as the volume of input data to the real time data warehouse increases, the interference between online loading ...
متن کامل