Using controlled vocabulary terms to optimize your search

In the previous video we learnt how to use search suggestions to help us choose effective keywords for our search. In this video, we’ll look at techniques to make those terms more precise.

A key task in developing effective search strategies is choosing the right terms for the various facets of your information need. For example, if you are researching the topic of promoting physical activity to prevent obesity in older people, you might start with a search like this:

Read Full Post »

Towards explainability in professional search

On Friday I was privileged to present a paper called ‘Towards Explainability in Professional Search‘ at the 3rd International Workshop on ExplainAble Recommendation and Search (EARS 2020), part of SIGIR 2020. This paper was co-presented with my colleague Andy MacFarlane of City University, and represents our collective thoughts and recommendations on how to develop more transparent, reproducible and explainable systems in professional search. Understandably, given our respective geographic locations the presentation was made remotely, and we are thankful for the local attendees who stayed around until 22:00 (local time) for our talk. The slides are available below. We view these recommendations as a conversation starter rather than the last word, so comments & feedback are particularly welcome. The paper itself is available for download from the EARS website.

Read Full Post »

How to optimize your search using query suggestions

In our previous video we learnt how to use visual approaches to create the correct structure for our search strategy. In this video, we’ll look at techniques to create the right content.

A key task in developing effective search strategies is choosing the right keywords. But how do we create these terms in the first place? One way is simply to brainstorm them, i.e. think up related terms for each facet. Or you might use your search results as a source of inspiration. But is there a more efficient way to generate related terms?

Read Full Post »

Putting search theory to work on large datasets

Last week I was privileged to present at the Singapore Chapter of ISKO on the topic of “Putting search theory to work on large datasets“. Understandably, given our respective geographic locations this presentation was made remotely, but I’m pleased to say we had a good attendance and a very informative discussion. Big thanks go to Patrick Lambe, Maish Nichani and Mark Garlinghouse for making this happen. The slides are available below, and the video should follow shortly. As always, comments & feedback welcome 🙂

Read Full Post »

How to optimize your search strategy using data visualization

In this video, we’ll look at how data visualization techniques can help us optimize and re-use search strategies.

The benefits offered by visualization become more apparent as the complexity of your search increases. For example, here is a relatively complex search string used by a recruiter to find social profiles for project managers in the Republic of Ireland. As you can see, it is quite hard to visualize how this search is structured, and even harder to debug or improve.

Read Full Post »

The benefits of visual searching

In our previous video we learnt how to build a search strategy and then save it and create a shareable link to it. In this video, we learn how to use visual approaches to make our searches smarter and more efficient.

Let’s return to the previous example where our research question concerned the role of physical activity in preventing obesity in older people. One of the immediate benefits of the 2D approach is that it you can see hit counts for each search block, displayed in the top right hand corner. You can use these to better understand which parts of your search to focus on when optimising your query.

Read Full Post »

Saving and sharing complex searches

In this short video we’re going to learn how to save, share and collaborate on structured searches.

Let’s return to the example we saw earlier, where our research question was to investigate promoting physical activity to prevent obesity in older people.

Choosing the correct terms can take some thought and effort but to save time in this video we have already entered our preferred terms on the canvas. To create the Boolean logic we saw earlier, we simply drag-and-drop terms to create groups:

Read Full Post »

2Dsearch: latest release including Lens.org and Yandex

It’s been a little while in the making, but I am delighted this week to announce a new release of 2Dsearch. This contains a variety of bug fixes and improvement, notably support for:

• Lens.org (a patent search facility and knowledge resource containing over 200m records)
• Yandex (5th largest search engine worldwide)

We’re particularly excited about the support for Lens.org, as it opens up all sorts of possibilities to find meta-analysis, systematic reviews, review articles of topics you are interested in regardless of your discipline.

Moreover, you can now use 2Dsearch to search visually across 8 different databases, and use automated translations for many more.  We’ve also made various improvements to the quality of the query parsing and the help/guidance provided on the messages tab, and added an in-app link to our tutorial videos. In addition, search strategies are now saved with an annotations field for your own notes and shared by name (rather than anonymously).

We’ve lots more planned for the next release, so if you’d like to help shape this and/or have comments of your own then do let us know. We’d be delighted to hear from you!

Read Full Post »

Intro to visual searching

In this short video we’re going to learn how to create a simple search strategy.

The best way to get started is to simply go to https://app.2dsearch.com and you’ll see a screen like this, with a drawing canvas on the left and a results pane on the right. At the moment our results pane is connected to Bing but you can see that other options are available. Note also that we can hide the results pane and show it at any time using this button. We also have various menu options available on the left which we’ll come to shortly.

To get started, we simply enter terms on the canvas. For example I can enter the terms hello and world and as I enter these terms, the results pane on the right updates dynamically. It starts to get interesting when we manipulate the objects on the canvas. For example, if I drag one term over the other it creates a new group with the default Boolean operator OR. Notice how the results pane on the right has again updated, showing this Boolean expression. If I want to change the operator I can simply edit it like this.

If I make a mistake I can just click the undo or redo buttons up here. Or I could just select an object and hit delete. I can create new terms or groups by copying and pasting like this. And I can add a negative term by entering it on the canvas and selecting the menu option apply not. If we want to see how our search looks as a traditional Boolean expression, we can click on the query tab.

In the next video, we’ll look at how to apply this to some real world searches.

Think outside the search box: https://www.2dsearch.com/

Read Full Post »

NLP techniques for automated query suggestions

A key task in formulating effective search strategies is the identification of appropriate keywords and controlled vocabulary terms. Interactive features such as query expansion can play a key role in supporting these tasks. In this presentation we investigate a variety of methods for interactive query expansion based on manually curated resources (e.g. ontologies and terminologies) and on distributional methods (e.g. unsupervised machine learning). The results demonstrate the utility of distributional models and the value of using ngram order to optimise precision and recall. This work was due to be presented at the ISKO UK event Using knowledge organization to deliver content (which will hopefully be rescheduled later this year).

Read Full Post »