Optimization of textual attribute support in generic location-aware rank query

Many information on the Internet nowadays is produced by mobile phone and becomes location-aware. Since the information has evolved to be more complex, a generic location-aware rank query problem was proposed (GLRQ). GLRQ allows searching the objects with different searchable attributes such as spatial, textual, and numeric together. To search many different types of attribute efficiently, a hybrid index structure called Inverted Files R-Tree with synopses tree (IRS) was proposed. However, the IRS was not able to achieve the optimized performance as the only spatial location is the only attribute that is taken into account in tree construction. Then Multi-IRS technique was proposed to generate multiple optimized trees for each numeric data type. However, the textual attribute was not handled. This paper presents the optimization technique for a textual attribute which can be included in Multi-IRS. The experimental results show that our proposed method reduces I/O access cost by 50% on the real dataset.