Hiroaki Ohshima

Trustworthiness analysis of web search results

Increased usage of Web search engines in our daily lives means that the trustworthiness of searched results has become crucial. User studies on the usage of search engines and analysis of the factors used to determine trust that users have in search results are described in this paper. Based on the analysis, we developed a system to help users determine the trustworthiness of Web search results by computing and showing each returned pages topic majority, topic coverage, locality of supporting pages (i.e., pages linked to each search result) and other information. The measures proposed in the paper can be applied to the search of Web-based libraries or can be useful in the usage of digital library search systems.

Searching coordinate terms with their context from the web

We propose a method for searching coordinate terms using a traditional Web search engine. “Coordinate terms” are terms which have the same hypernym. There are several research methods that acquire coordinate terms, but they need parsed corpora or a lot of computation time. Our system does not need any preprocessing and can rapidly acquire coordinate terms for any query term. It uses a conventional Web search engine to do two searches where queries are generated by connecting the users query term with a conjunction “OR”. It also obtains background context shared by the query term and each returned coordinate term.

Mining the web for hyponymy relations based on property inheritance

Concept hierarchies, such as hyponymy and meronymy relations, are very important for various natural language processing systems. Many researchers have tackled how to mine very large corpora of documents such as the Web for them not manually but automatically. However, their methods are mostly based on lexico-syntactic patterns as not necessary but sufficient conditions of concept hierarchies, so they can achieve high precision but low recall when using stricter patterns or they can achieve high recall but low precision when using looser patterns. In this paper, property inheritance from a concept to its hyponyms is assumed to be necessary and sufficient conditions of hyponymy relations to achieve high recall and not low precision, and we propose a method to acquire hyponymy relations from the Web based on property inheritance.

What can history tell us?: towards different models of interaction with document histories

The current Web is a dynamic collection where little effort is made to version pages or to enable users to access historical data. As a consequence, they generally do not have sufficient temporal support when browsing the Web. However, we think that there are many benefits to be obtained from integrating documents with their histories. For example, a documents history can enable us to travel back through time to establish its trustworthiness. This paper discusses the possible types of interactions that users could have with document histories and it presents several examples of systems that we have implemented for utilizing this historical data. To support our view, we present the results of an online survey conducted with the objective of investigating user needs for temporal support on the Web. Although the results indicated quite low use of Web archives by users, they simultaneously emphasized their considerable interest in page histories.

Can Social Tagging Improve Web Image Search

Conventional Web image search engines can return reasonably accurate results for queries containing concrete terms, but the results are less accurate for queries containing only abstract terms, such as spring or peace. To improve the recall ratio without drastically degrading the precision ratio, we developed a method that replaces an abstract query term given by a user with a set of concrete terms and that uses these terms in queries input into Web image search engines. Concrete terms are found for a given abstract term by making use of social tagging information extracted from a social photo sharing system, such as Flickr. This information is rich in user impressions about the objects in the images. The extraction and replacement are done by (1) collecting social tags that include the abstract term, (2) clustering the tags in accordance with the term co-occurrence of images, (3) selecting concrete terms from the clusters by using WordNet, and (4) identifying sets of concrete terms that are associated with the target abstract term by using a technique for association rule mining. Experimental results show that our method improves the recall ratio of Web image searches.

Context matcher: improved web search using query term context in source document and in search results

When reading a Web page or editing a word processing document, we often search the Web by using a term on the page or in the document as part of a query. There is thus a correlation between the purpose for the search and the document being read or edited. Modifying the query to reflect this purpose can thus improve the relevance of the search results. There have been several attempts to extract keywords from the text surrounding the search term and add them to the initial query. However, identifying appropriate additional keywords is difficult; moreover, existing methods rely on precomputed domain knowledge. We have developed Context Matcher: a query modification method that uses the text surrounding the search term in the initial search results as well as the text surrounding the term in the document being read or edited, the “source document”. It uses the text surrounding the search term in the initial results to weight candidate keywords in the source document for use in query modification. Experiments showed that our method often found documents more related to the source document than baseline methods that use context either in only the source document or search results.

Visualizing Changes in Coordinate Terms over Time: An Example of Mining Repositories of Temporal Data through their Search Interfaces

Certain data repositories provide search functionality for temporally ordered data. News archive search or blog search are examples of search interfaces that allow issuing structured queries composed of arbitrary terms and selected time constraints for performing temporal search. However, extracting aggregated knowledge such as detecting the evolution of objects or their relationships through these interfaces is difficult for users. In this paper, we discuss the problem of knowledge extraction and agglomeration from repositories of temporal data. In particular, we propose a method for detecting and visualizing changes in coordinate terms over time based on a news archive.

Sibling page search by page examples

We propose methods of searching Web pages that are “semantically” regarded as “siblings” with respect to given page examples. That is, our approach aims to find pages that are similar in theme but have different content from the given sample pages. We called this “sibling page search”. The proposed search methods are different from conventional content-based similarity search for Web pages. Our approach recommends Web pages whose “conceptual” classification category is the same as that of the given sample pages, but whose content is different from the sample pages. In this sense, our approach will be useful for supporting a users opportunistic search, meaning a search in which the users interest and intention are not fixed. The proposed methods were implemented by computing the “common” and “unique” feature vectors of the given sample pages, and by comparing those feature vectors with each retrieved page. We evaluated our method for sibling page search, in which our method was applied to test sets consisting of page collections from the Open Directory Project (ODP).

ReCQ: real-world context-aware querying

This paper proposes a method of context-aware querying in mobile/ubiquitous Web searches, which provides mobile users with four capabilities: (1) context-aware keyphrase inference to help them input a keyphrase as a part of their keyword-based query, (2) context-aware subtopic tree generation to help them specify their information demand on one subtopic, (3) discovery of comparable keyphrases to their original query to help them make better decisions, and (4) meta vertical search focused on one subtopic to make the retrieval results more precise.

Query Relaxation Based on Users' Unconfidences on Query Terms and Web Knowledge Extraction

This paper proposes a method to allow users to search for Web pages according to their search intentions. We introduce a degree of unconfidence for each term in a Web search query. We first investigate the relationships among query terms by accessing a Web search engine. Next, according to a degree of users unconfidences for each query term and the relationships among query terms, our system finds alternative terms by accessing to a Web search engine. Then, our system generates a collection of new queries that are different from the original query and merges the Web search results obtained for each new query. We implemented our system and showed the usefulness of our system based on user evaluation.