Название: Hybrid Intelligent Systems for Information Retrieval
Автор: Anuradha D. Thakare, Shilpa Laddha, and Ambika Pawar
Издательство: CRC Press
Год: 2023
Страниц: 253
Язык: английский
Формат: pdf (true)
Размер: 10.1 MB

Hybrid Intelligent Systems for Information Retrieval covers three areas along with the introduction to Intelligent IR, i.e., Optimal Information Retrieval Using Evolutionary Approaches, Semantic Search for Web Information Retrieval, and Natural Language Processing for Information Retrieval.

This book covers the architectures of modern information systems pertaining to structured and unstructured data retrieval, and there is a detailed discussion on how to develop computational models for retrieval systems. It describes evolutionary approaches for optimal information retrieval and the design of hybrid intelligent information retrieval systems for various applications. The focus is on three key areas: Optimality in Information Retrieval with Evolutionary Algorithms, Semantic Web Information Retrieval, and Natural Language Processing for Information Retrieval. This is a comprehensive textbook on the subject, covering a broad array of topics with more emphasis on the techniques that have been profitably employed in exploiting the available information. To give a clear understanding of the topics, the case studies and examples of hybrid intelligent information systems are also included.

Deep learning methods have proven applicability in information retrieval (IR). Deep learning models eliminate human bias for feature or relevance measure and make it more efficient. Deep learning has a lot of potential to improvise IR. Recurrent neural networks (RNNs) are highly efficient with an internal memory, and it has the most promising algorithms. RNN has gained more popularity with enhanced computation capacity, large volume of data, and long short-term memory that is known as LSTM. In sequen­ tial data, things are in order and follow each other, e.g., DNA sequence data and time-series data. RNN stores input in internal memory; there­ fore, it is more useful for problems with sequential data. Also, RNN stores important information about input in internal memory, thus improving accuracy in predicting the next stage. RNN algorithms are applicable to sequential data because of internal memory storage capacity, e.g., it can be efficiently used for time series, text, financial data, and so on. With respect to other algorithms, RNN algorithms have a very deep understanding of a sequence and about context. RNNs provide more accurate predictive results for input sequential data than other algorithms.

- Talks about the design, implementation, and performance issues of the hybrid intelligent information retrieval system in one book
- Gives a clear insight into challenges and issues in designing a hybrid information retrieval system
- Includes case studies on structured and unstructured data for hybrid intelligent information retrieval
- Provides research directions for the design and development of intelligent search engines

This book is aimed primarily at graduates and researchers in the information retrieval domain.