Um método em dois níveis para complementação automática de sentenças

Abstract

Query autocompletion has become a standard feature in many applications that use textual commands, especially for search engines, significantly increasing the quality of the experience of using these applications. By observing the automatic complementation methods present in the literature, the main factors that indicate the viability of these methods for certain applications are the query process time and the amount of memory used to index the data.

This work presents a new method for query autocompletion that performs the search in two levels, allowing a significant saving of memory space, maintaining a performance in query processing time acceptable in relation to the main methods present in the literature. Experiments performed in databases of different sizes indicate that the presented method significantly reduces the amount of memory required to perform automatic complementation when compared to published works. This result is obtained by maintaining an acceptable processing time and, in some scenarios, can be even better than those obtained by the main works found in the literature.