Saturday July 04 2009 | Biotechnology feed | All feeds

BioPortfolio web search engine Research Report March 2001

By Silico Research ( www.Silico-Research.com )

Executive summary Product BioPortfolio search engine Date 2nd March, 2001
Background. Silico Research has been engaged by Bioportfolio to conduct an independent analysis of Bioportfolio’s web search engine ( http://www.bioportfolio.co.uk ). The search engine is designed to be used by pharmaceutical and healthcare executives and scientists for sector-specific searches. The search engine is built upon SmartLogik™ technology developed Muscat. Muscat is a subsidiary of Bright Station, a UK-based company.

Bioportfolio takes a different approach from other search engines. General web sites typically index all sites submitted to them. Bioportfolio’s search engine indexes and classifies 4,000 sector-relevant web sites selected by Bioportfolio’ staff. Bioportfolio plans to increase the number of sites indexed to 6,000 this year.

Bioportfolio also differs from other search engines in that it uses a pharmaceutical and scientific thesaurus to classify and index web-pages. General search engines typically index web pages without reference to a thesaurus or taxonomy. The taxonomy is used to suggest words to add to the search and to allow the user to refine his search.

Methodology. For the purposes of our analysis we conducted four word-based searches on Bioportfolio, Google (http://www.google.com) and Northern Light (http://www.northernlight.com). Google and Northern Light were used as comparisons because the evidence is that they are among the most heavily used search engines by the scientific and pharmaceutical communities. Search words included ‘proteome’, ‘bioinformatics’, ‘data-mining’ and ‘BLAST’.

Conclusion. We believe that the Bioportfolio search engine is one of the most valuable tools available to pharmaceutical executives and scientists who search the web on a regular basis. We found that Bioportfolio typically generated search results that are likely to be of a higher value to biopharmaceutical executives and scientists than either Google or Northern Light. There were fewer expired pages in the Bioportfolio search and fewer low value pages. We classified pages as low value where, for example, they were student-orientated, were predominately advertising-related or were themselves directories of other web pages.

Executive summary of findings
Bioportfolio main beneficial features
We found that Bioportfolio had a number of features that we believe makes it a more useful search engine for pharmaceutical and scientific users than Google or Northern Light.

Document types. Bioportfolio indexes html pages, PDF files, MS PowerPoint files and MS Word documents. Whilst it has recently begun to index PDF files Google does not index PowerPoint files or Word documents. Northern Light does not index PDF files or MS Word documents. As many pharmaceutical and scientific web sites include PowerPoint presentations or Word documents the ability to access both is a very useful feature of Bioportfolio’s search engine. On average 8% of the documents returned by Bioportfolio in our test searches were in PowerPoint or Word formats.
Stemming. Bioportfolio searches for the stem of the words entered. So, for example, entering the word ‘bioinformatic’ returns a result based on the occurrences of the stem ‘bioinformat’. Google and Northern Light search strictly on the word entered.
Searching. Both Bioportfolio and Northern Light return a probabilistic result based upon the occurrence of the words entered. This means that a number of words and alternatives can be entered in order to generate a more precise search.
Suggestions. Bioportfolio suggests additional words to search for. So, for example, a search for ‘bioinformatic’ and ‘BLAST’ will return a number of additional terms to search on including ‘waterman’, ‘genbank’, ‘fasta’, ‘proteom’, ‘smith’ and ‘genom’. We found this a very useful feature for refining searches.
Dating. Both Bioportfolio and Northern Light return a date for the document and allow for searches to be limited and sorted by date. Google does not have this feature.
Highlighting. Bioportfolio highlights the occurrences of the search terms in the returned documents. This is a useful feature in long or complex documents.

Suggestions for improvement
We believe that the Bioportfolio service would benefit from a number of features:

Accessibility. We would recommend that Bioportfolio seriously considers reducing the number of screens that need to be passed through in order to access the service. The user passers through four screens including a password screen in order to execute a search. This is likely to act as a deterrent to many users and so to reduce the usage of the service.
Site addition. We would suggest that Bioportfolio considers a service allowing site owners to submit their site for inclusion on the service. This is standard in other search engines and would aid the task of making the service comprehensive. We believe that there will be a high level of participation and a low percentage of irrelevant sites because of the specialised nature of the users of the service.
Hyperlinks. We would suggest that Bioportfolio examines whether it is possible automatically to add sites to the database that are hyper-linked in the indexed sites. This would significantly increase the number of sites in the database without necessarily reducing the quality of the sites.

Help files:

Nothing in this website should be used in place of personal medical advice from your own qualified medical practitioner.