Inverted index is used in most Information Retrieval Systems (IRS) to achieve the fast query response time. In inverted index, compression schemes are used to improve the efficiency of IRS. In this chapter, the authors study and analyze various compression techniques that are used for indexing. They also present a new compression technique that is based on FastPFOR called New FastPFOR. The storage structure and the integers' representation of the proposed method can improve its performances both in compression and decompression. The study on existing works shows that the recent research works provide good results either in compression or in decoding, but not in both. Hence, their decompression performance is not fair. To achieve better performance in decompression, the authors propose New FastPFOR in this chapter. To evaluate the performance of the proposed method, they experiment with TREC collections. The results show that the proposed method could achieve better decompression performance than the existing techniques.
Techniques for Inverted Index Compression
