Jaana Kekäläinen

Cumulated gain-based evaluation of IR techniques

Modern large retrieval environments tend to overwhelm their users by their large output. Since all documents are not of equal relevance to their users, highly relevant documents should be identified and ranked first for presentation. In order to develop IR techniques in this direction, it is necessary to develop evaluation approaches and methods that credit IR methods for their ability to retrieve highly relevant documents. This can be done by extending traditional evaluation methods, that is, recall and precision based on binary relevance judgments, to graded relevance judgments. Alternatively, novel measures based on graded relevance judgments may be developed. This article proposes several novel measures that compute the cumulative gain the user obtains by examining the retrieval result up to a given ranked position. The first one accumulates the relevance scores of retrieved documents along the ranked result list. The second one is similar but applies a discount factor to the relevance scores in order to devaluate late-retrieved documents. The third one computes the relative-to-the-ideal performance of IR techniques, based on the cumulative gain they are able to yield. These novel measures are defined and discussed and their use is demonstrated in a case study using TREC data: sample system run results for 20 queries in TREC-7. As a relevance base we used novel graded relevance judgments on a four-point scale. The test results indicate that the proposed measures credit IR methods for their ability to retrieve highly relevant documents and allow testing of statistical significance of effectiveness differences. The graphs based on the measures also provide insight into the performance IR techniques and allow interpretation, for example, from the user point of view.

IR evaluation methods for retrieving highly relevant documents

This paper proposes evaluation methods based on the use of non-dichotomous relevance judgements in IR experiments. It is argued that evaluation methods should credit IR methods for their ability to retrieve highly relevant documents. This is desirable from the user point of view in modem large IR environments. The proposed methods are (1) a novel application of P-R curves and average precision computations based on separate recall bases for documents of different degrees of relevance, and (2) two novel measures computing the cumulative gain the user obtains by examining the retrieval result up to a given ranked position. We then demonstrate the use of these evaluation methods in a case study on the effectiveness of query types, based on combinations of query structures and expansion, in retrieving documents of various degrees of relevance. The test was run with a best match retrieval system (In- Query I) in a text database consisting of newspaper articles. The results indicate that the tested strong query structures are most effective in retrieving highly relevant documents. The differences between the query types are practically essential and statistically significant. More generally, the novel evaluation methods and the case demonstrate that non-dichotomous relevance assessments are applicable in IR experiments, may reveal interesting phenomena, and allow harder testing of IR methods.

Using graded relevance assessments in IR evaluation

This article proposes evaluation methods based on the use of nondichotomous relevance judgements in IR experiments. It is argued that evaluation methods should credit IR methods for their ability to retrieve highly relevant documents. This is desirable from the user point of view in modern large IR environments. The proposed methods are (1) a novel application of P-R curves and average precision computations based on separate recall bases for documents of different degrees of relevance, and (2) generalized recall and precision based directly on multiple grade relevance assessments (i.e., not dichotomizing the assessments). We demonstrate the use of the traditional and the novel evaluation measures in a case study on the effectiveness of query types, based on combinations of query structures and expansion, in retrieving documents of various degrees of relevance. The test was run with a best match retrieval system (InQuery1) in a text database consisting of newspaper articles. To gain insight into the retrieval process, one should use both graded relevance assessments and effectiveness measures that enable one to observe the differences, if any, between retrieval methods in retrieving documents of different levels of relevance. In modern times of information overload, one should pay attention, in particular, to the capability of retrieval methods retrieving highly relevant documents.

The impact of query structure and query expansion on retrieval performance

The effects of query structures and query expansion (QE) on retrieval performance were tested with a best match retrieval system (INQUERY1). Query structure means the use of operators to express the relations between search keys. Eight different structures were tested, representing weak structures (averages and weighted averages of the weights of the keys) and strong structures (e.g., queries with more elaborated search key relations). QE was based on concepts, which were first selected from a conceptual model, and then expanded by semantic relationships given in the model. The expansion levels were (a) no expansion, (b) a synonym expansion, (c) a narrower concept expansion, (d) an associative concept expansion, and (e) a cumulative expansion of all other expansions. With weak structures and Boolean structured queries, QE was not very effective. The best performance was achieved with one of the strong structures at the largest expansion level.

Binary and graded relevance in IR evaluations-Comparison of the effects on ranking of IR systems

In this study the rankings of IR systems based on binary and graded relevance in TREC 7 and 8 data are compared. Relevance of a sample TREC results is reassessed using a relevance scale with four levels: non-relevant, marginally relevant, fairly relevant, highly relevant. Twenty-one topics and 90 systems from TREC 7 and 20 topics and 121 systems from TREC 8 form the data. Binary precision, and cumulated gain, discounted cumulated gain and normalised discounted cumulated gain are the measures compared. Different weighting schemes for relevance levels are tested with cumulated gain measures. Kendalls rank correlations are computed to determine to what extent the rankings produced by different measures are similar. Weighting schemes from binary to emphasising highly relevant documents form a continuum, where the measures correlate strongly in the binary end, and less in the heavily weighted end. The results show the different character of the measures.

ExpansionTool: Concept-Based Query Expansion and Construction

We develop a deductive data model for concept-based query expansion. It is based on three abstraction levels: the conceptual, linguistic and string levels. Concepts and relationships among them are represented at the conceptual level. The linguistic level gives natural language expressions for concepts. Each expression has one or more matching patterns at the string level. The models specify the matching of the expression in database indices built in varying ways. The data model supports a declarative concept-based query expansion and formulation tool, the ExpansionTool, for heterogeneous IR system environments. Conceptual expansion is implemented by a novel intelligent operator for traversing transitive relationships among cyclic concept networks. The number of expansion links followed, their types, and weights can be used to control expansion. A sample empirical experiment illustrating the use of the ExpansionTool in IR experiments is presented.

The polyrepresentation continuum in IR

The polyrepresentation principle suggests that cognitively and functionally different representations of information objects may be used in information retrieval to enhance quality of results. In the paper, several empirical studies that intentionally or unintentionally have tested the principle are introduced and discussed. The continuum proposed by Larsen (2004; Ingwersen & Larsen, 2005) showing the structural dimension of the retrieval techniques involved in polyrepresentation is further elaborated by adding a novel second dimension consisting of query structure and modus. The new two-dimensional continuum can be seen as a constructive framework for further investigations of polyrepresentative principles in IR.

Generalized contextualization method for XML information retrieval

A general re-weighting method, called contextualization, for more efficient element ranking in XML retrieval is introduced. Re-weighting is based on the idea of using the ancestors of an element as a context: if the element appears in a good context -- good interpreted as probability of relevance -- its weight is increased in relevance scoring; if the element appears in a bad context, its weight is decreased. The formal presentation of contextualization is given in a general XML representation and manipulation frame, which is based on utilization of structural indices. This provides a general approach independent of weighting schemas or query languages.Contextualization is evaluated with the INEX test collection. We tested four runs: no contextualization, parent, root and tower contextualizations. The contextualization runs were significantly better than no contextualization. The root contextualization was the best among the re-weighted runs.

Document text characteristics affect the ranking of the most relevant documents by expanded structured queries

The increasing flood of documentary information through the Internet and other information sources challenges the developers of information retrieval systems. It is not enough that an IR system is able to make a distinction between relevant and non‐relevant documents. The reduction of information overload requires that IR systems provide the capability of screening the most valuable documents out of the mass of potentially or marginally relevant documents. This paper introduces a new concept‐based method to analyse the text characteristics of documents at varying relevance levels. The results of the document analysis were applied in an experiment on query expansion (QE) in a probabilistic IR system. Statistical differences in textual characteristics of highly relevant and less relevant documents were investigated by applying a facet analysis technique. In highly relevant documents a larger number of aspects of the request were discussed, searchable expressions for the aspects were distributed over a larger set of text paragraphs, and a larger set of unique expressions were used per aspect than in marginally relevant documents. A query expansion experiment verified that the findings of the text analysis can be exploited in formulating more effective queries for best match retrieval in the search for highly relevant documents. The results revealed that expanded queries with concept‐based structures performed better than unexpanded queries or Nnatural languageO queries. Further, it was shown that highly relevant documents benefit essentially more from the concept‐based QE in ranking than marginally relevant documents.

The Co-Effects of Query Structure and Expansion on RetrievalPerformance in Probabilistic Text Retrieval

The effects of query structures and query expansion (QE) on retrieval performance were tested with a best match retrieval system (InQuery1). Query structure means the use of operators to express the relations between search keys. Six different structures were tested, representing strong structures (e.g., queries with facets or concepts identified) and weak structures (no concepts identified, a query is ‘a bag of search keys’). QE was based on concepts, which were first selected from a searching thesaurus, and then expanded by semantic relationships given in the thesaurus. The expansion levels were (a) no expansion, (b) a synonym expansion, (c) a narrower concept expansion, (d) an associative concept expansion, and (e) a cumulative expansion of all other expansions. With weak structures and Boolean structured queries, QE was not very effective. The best performance was achieved with a combination of a facet structure, where search keys within a facet were treated as instances of one search key (the SYN operator), and the largest expansion.