AdaGP-Rank: Applying boosting technique to genetic programming for learning to rank

The field of information retrieval has witnessed tangible progress over the past decades in response to the expanded usage of the internet and the dire need of users to search for massive amounts of digital information. Given the steady increase of Arabic e-content, excellent information retrieval systems must be devised to suit the nature and requirements of the Arabic language. This paper sheds light on the current progress in the field of Arabic information retrieval, identifies the challenges that hinder the progress of this science, and proposes suggestions for further research. This paper uses the descriptive analytical method to examine the reality of Arabic studies in the field of information retrieval and to study the problems that are being faced in this area. Specifically, the previous literature on information retrieval is reviewed by searching the related databases and websites.

The document-length normalization problem has been widely studied in the field of information retrieval. The cosine normalization (Baeza-Yates and Ribeiro-Neto, 1999), the maximum if normalization (Allan et al., 1997) and the byte length normalization (Robertson et al., 1992) are the most commonly used normalization techniques. In (Singhal et al., 1996), authors studied the retrieval probability of documents w.r.t. their size, using different similarity measures. They have shown that none of existing measures retrieve the documents of different lengths with the same probability. We first show here that the document and query sizes are indeed very influent on the similarity score expectation. Therefore, we propose to realize a statistical regression of the similarity scores distribution w. r. t. document and query sizes in order to normalize them. Experimental results appear to indicate that our approach, as well in the field of classical Information Retrieval as when applied to a document clustering process, allows to judge similarities really more fairly.

Semi-supervised ranking is a relatively new and important learning problem. Some semi-supervised learning to rank algorithms have been proposed by adapting existing pair-wise ranking algorithms into the semi-supervised setting. In contrast, there is comparatively little work in developing semi-supervised ranking algorithm based on the list-wise approach. In this paper, we proposed a semi-supervised extension of one of the state-of-the art list-wise ranking algorithm--AdaRank, for the task of learning to rank with partially labeled data. The proposed algorithm is a two-staged approach that combines the process of label propagation and a regularized version of AdaRank. The label propagation process is used to propagate the relevance labels from labeled instances to other unlabeled ones, so that more training data are available to learn the ranking function. The regularized version of AdaRank optimizes a novel performance measure, which incorporates both the usual Information Retrieval IR metrics and the smoothness of the ranking function. Experimental results on benchmark datasets show the advantage of our methods over existing pair-wise algorithms for the Information Retrieval measures.

A generic ranking function discovery framework by genetic programming for information retrieval

Recently, ranking data with respect to the intrinsic geometric structure (manifold ranking) has received considerable attentions, with encouraging performance in many applications in pattern recognition, information retrieval and recommendation systems. Most of the existing manifold ranking methods focus on learning a ranking function that varies smoothly along the data manifold. However, beyond smoothness, a desirable ranking function should vary monotonically along the geodesics of the data manifold, such that the ranking order along the geodesics is preserved. In this article, we aim to learn a ranking function that varies linearly and therefore monotonically along the geodesics of the data manifold. Recent theoretical work shows that the gradient field of a linear function on the manifold has to be a parallel vector field. Therefore, we propose a novel ranking algorithm on the data manifolds, called Parallel Field Ranking. Specifically, we try to learn a ranking function and a vector field simultaneously. We require the vector field to be close to the gradient field of the ranking function, and the vector field to be as parallel as possible. Moreover, we require the value of the ranking function at the query point to be the highest, and then decrease linearly along the manifold. Experimental results on both synthetic data and real data demonstrate the effectiveness of our proposed algorithm.

Document analysis is one of the emerging area of research in the field of Information Retrieval. Many attempts have been made for retrieving information from a document using various machine learning algorithms. A concept of context vector is frequently used in information retrieval from document/s. Context Vector is an vector, which is used for various feature selection from documents, automatic classification of text documents, Subject Verb Agreement, etc. This paper discusses, the attempts made in the field of Information Retrieval (IR) from document using context vector. It also discuss about pros and cons of each attempt. This paper propose a system which can give “context vector” of the document set using Latent Semantic Analysis which is the most trending method in document analysis. The system is tested on BBC news dataset and proves to be successful.

Queries describe the users' search intent and therefore they play an essential role in the context of ranking for information retrieval and Web search. However, most of existing approaches for ranking do not explicitly take into consideration the fact that queries vary significantly along several dimensions and entail different treatments regarding the ranking models. In this paper, we propose to incorporate query difference into ranking by introducing query-dependent loss functions. In the context of Web search, query difference is usually represented as different query categories; and, queries are usually classified according to search intent such as navigational, informational and transactional queries. Based on the observation that such kind of query categorization has high correlation with the user's different expectation on the result accuracy on different rank positions, we develop position-sensitive query-dependent loss functions exploring such kind of query categorization. Beyond the simple learning method that builds ranking functions with pre-defined query categorization, we further propose a new method that learns both ranking functions and query categorization simultaneously. We apply the query-dependent loss functions to two particular ranking algorithms, RankNet and ListMLE. Experimental results demonstrate that query-dependent loss functions can be exploited to significantly improve the accuracy of learned ranking functions. We also show that the ranking function jointly learned with query categorization can achieve better performance than that learned with pre-defined query categorization. Finally, we provide analysis and conduct additional experiments to gain deeper understanding on the advantages of ranking with query-dependent loss functions over other query-dependent ranking approaches and query-independent approaches.

This article discusses the opportunities and challenges of applying modern information retrieval techniques to the cultural heritage domain. Although the field of information retrieval is closely associated with computer science, it originally emerged from library science — also one of the main disciplines concerned with access to cultural heritage material. Hence we are, in a sense, exploring what happens if we bring these strands of research back together again. The article consists of three parts. In the first part, we explain the field of information retrieval and its multidisciplinary nature. In the second part, we discuss how and why the problem of providing access to cultural heritage can be cast naturally as an information retrieval problem. In the third and main part, we present a detailed case study of applying the modern information retrieval approach in practice within a museum.

Ranking is a core problem for information retrieval since the performance of the search system is directly impacted by the accuracy of ranking results. Ranking model construction has been the focus of both the fields of information retrieval and machine learning, and learning to rank in particular has attracted much interest. Many ranking models have been proposed, for example, RankSVM is a state-of-the-art method for learning to rank and has been empirically demonstrated to be effective. However, most of the proposed methods do not consider about the significant differences between queries, only resort to a single function in ranking. In this paper, we present a novel ranking model named QoRank, which performs the learning task dependent on queries. We also propose a LSE (least-squares estimation) -based weighted method to aggregate the ranking lists produced by base decision functions as the final ranking. Comparison of QoRank with other ranking techniques is conducted, and several evaluation criteria are employed to evaluate its performance. Experimental results on the LETOR OHSUMED data set show that QoRank strikes a good balance of accuracy and complexity, and outperforms the baseline methods. © 2010 Wiley Periodicals, Inc.

In recent years, information retrieval technology is widely used in the medical industry. How to effectively retrieve electronic medical record has become a hot issue in the field of information retrieval. Medical terms occupy an important position in the electronic medical record (EMR) retrieval, and they are usually used to limit the retrieval conditions, so they suggest the user’s search intention. Aiming at the importance of medical terms, a method of query reformulation based on medical term reweighting is proposed to improve the performance of EMR retrieval. Firstly, the method filtrates medical terms from the origin query. Then, each medical term is weighted by its own self-information which can be estimated from the document set. Finally, the new query is constructed by combining the weighted medical terms and origin query proportionally. Experimental results on the TREC dataset show that our method performed above the baseline in three performance metrics: MAP (+ 14.2%), bpref (+ 8.0%), and P@10 (+ 9.6%).

Retrieval of relevant documents from a collection is a tedious task. As genetic algorithms (GA) are robust and efficient search and optimization techniques, they can be used to search the huge document search space. In this paper, a general frame work of information retrieval system is discussed. The applicability of genetic algorithms in the field of information retrieval is also discussed. A review on how GA is applied to different problem domains in information retrieval is presented. A study on the similarity measures in information retrieval is also discussed.

Research on information retrieval model under scarcity theory and user cognition

This paper presents a survey of evolutionary approach for Clustering and how Genetic Algorithm is applied to different problem domains in the field of Information Retrieval. The paper is original in two main aspects first it gives an overview of evolutionary approach for clustering and information retrieval and second it suggests a novel evolutionary approach which serves both the areas of Clustering and Information Retrieval by developing a single Genetic Algorithm. Conventional clustering algorithms are sensitive to the choice of cluster center. Such a partition is insufficient to represent many real situations. To improve the performance of clustering algorithm, Genetic Algorithms is applied to clustering algorithms. Genetic Algorithms are robust in searching a multidimensional space to find optimal solution. This paper tries to reflect the research profile of this area by focusing more on how Evolutionary Algorithms can effectively work in forming clusters and improving retrieval performance.

The document retrieval is one of the fast growing and complex research area in the field of information retrieval. Concept-based information retrieval is more beneficial than keyword-based access. The most important advantage is the ability to make use of semantic relationships among concepts in finding relevant documents. Removal of irrelevant documents by identifying conceptual mismatches is another benefit. The linguistic representation of concepts is words and phrases. These concepts are mental structures. The inherent conciseness of natural language can make words represent multiple concepts. Different words can represent the same or similar concepts. We present a novel method of concept mining methodology, which provides an efficient way of document retrieval. The performance of this proposed retrieval system was analyzed by the existing methods for different medical documents. The proposed retrieval system shows that the documents are retrieved effectively based on concepts and the system performs with higher accuracy value.

Since its inception in the late 1950s, the field of Information Retrieval (IR) has developed tools that help people find, organize, and analyze information. The key early influences on the field are well-known. Among them are H. P. Luhn's pioneering work, the development of the vector space retrieval model by Salton and his students, Cleverdon's development of the Cranfield experimental methodology, Spärck Jones' development of idf , and a series of probabilistic retrieval models by Robertson and Croft. Until the development of the WorldWideWeb (Web), IR was of greatest interest to professional information analysts such as librarians, intelligence analysts, the legal community, and the pharmaceutical industry.