This guide will give you some ideas about how to improve your searches. It assumes that you have connectors and search language at your fingertips. If you need to, please go and look at the advanced search guide to find out how Vable search language differs from other databases.
Given that you can add a number of searches to a Topic, you can mix up individual search concepts - keeping them relatively simple or making them more complex. In this way you can cover everything. You can also reuse and repurpose searches by cloning the Topic and amending it, as required.
Keywords and Field Searching
IBM Watson processes the available text of all English language articles coming into Vable InfoPro. Using the words in the text, keywords are extracted. This information can be used in a number of ways:
- Companies and Location
- Industry and Subjects
- Keywords search
Industry and Subjects are maintained and monitored by the data team at Vable. We keep an eye on these so that we include new keywords in the Taxonomy. As we will see with some terms, there are always new concepts, developments, synonyms and acronyms to be considered.
The last one - keywords - is the most useful for advanced searching. For example if you type “keywords:pharma*”, Vable will find pharmaceutical, big pharma etc in the keywords associated with that article. If you want to find more than one keyword, type keywords:(pharma* OR “drug manufact*)
Searching across specific fields is a great way of focusing your searches, and you have two options for field searching. For an instant field search use the dropdown menu which appears when you type within the search box, which gives you a choice of URL, title, full text, intro text or summary. Use summary with care as publishers don’t always include any text in it.
Or you can utilize the field searches by typing in a command. You have the same field choices as in the dropdown menu, with the benefit of being able to combine them. For example title:ethics AND url:hbr*
- Keep the search broad and use * when limiting to fields.
- Always include a * after the URL term.
- Always check how the command should be formatted, eg URL needs to be url:, intro text has to be introText: etc.
Worked Example 1
(law OR legal) AND (tech* OR ("artificial intelligence" OR AI))
This search is correct and groups together the appropriate concepts, but the sheer volume of results makes it less than useful. So how can we improve this?
- We could create a custom filter of sources or utilize a private taxonomy. If we wanted to focus on a specific group of relevant publishers or publications, that’s a great start. If you are interested in tech sources there are some listed on our blog here.
However one of the benefits of Vable is finding articles in publications we wouldn't necessarily consider, so sometimes we don't want to limit by source. As a result we need to amend the search.
- Law and legal are great, but rather generic terms. We want to find articles specific to the legal industry.
- The truncated Tech* is a difficult term especially when it is combined with AI. The resulting search is too broad. However we could use AND.
- AI/”artificial intelligence” is fine and the focus of the search. It’s good to include the acronym or shortened version, with its longer term
Over the 7 days, the results are reduced from 13,000 to 50 using this search string:
keywords:(law OR legal) AND introText:(tech* AND ("artificial intelligence" OR AI))
Worked Example 2
introText:"artificial intelligence" AND ("law firm*" OR "knowledge management")
To stay on top of news about artificial intelligence in legal knowledge management, this search is correct, suitably narrow, and has the appropriate concepts. But how can we improve this?
- Again we could limit by source.
- “Law firm” could be broadened to include legal industry and law profession. This could be achieved by limiting it to “law OR legal” in the keywords.
- "Knowledge management" could be expanded to include KM.
- ”Artificial intelligence” is fine and the focus of the search. It’s good to include the acronym or shortened version, with its longer term
The following search is broader without being unwieldy, and is still running across the entire database:
("artificial intelligence" OR AI) AND ("knowledge management" OR KM) AND keywords:(law OR legal)
Worked Example 3
If you need to stay on top of AI in the legal world, you might want a general overview search to feed into a Topic.
- Articles should be from across the entire legal industry (law OR legal OR judiciary OR justice)
- "Artificial intelligence", although recognized as a broad term, is a troublesome and a lot of articles also refer (rightly) to the specific applications of AI.
- Many articles now have useful terms in the title so although you might miss some, you will bring back many relevant results.
- Don't forget you can exclude words or use/exclude filters
The following search is broad and runs across the entire database but focuses just on the titles of articles:
title:(AI OR ML OR "artificial intelligence" OR "machine learning" OR robotics OR dsml OR "big data" OR "data science" OR "machine intelligence" OR NLP OR "natural language process*" OR legaltech OR "start-up*" OR lawtech) AND keywords:(law OR legal OR judiciary OR justice) -police -war
For a fully effective AI-based Topic, this search will need to be supplemented by other more specific source feed searches, as well as specific searches with a filter applied. For example, "machine learning" with a source filter applied.