Applying Context Discovery to the Election Debate Mind Maps

In my previous two posts you can see mind maps summarising the content of the UK Election TV Debates between the main party leaders Gordon Brown, David Cameron and Nick Clegg.  I have been waiting to find an application for Context Discovery with my online and one chinese meal in Soho friend, Henry Lewkowicz.  These maps seem ideal there is lots of content and it’s not easy to get an objective overview.  I installed Context Discovery Mindmanager option and this is what happened.

I exported the MindManager maps to MS Word. You can’t run Context Discovery on the map, it operates on documents and web pages linked to the map.  There is a new set of Context Discovery commands in the Add ins toolbar of MindManager which allow you to set the parameters such as how many Top Keywords and whether to include summaries.  You then select the linked topic, click the Insert Content command and the magic happens.

I thought there was some overlap between the two debates but the Context Discovery map seems to say otherwise.  Only “jobs” is in the top ten of both debates.  Note: the maps are my precis of their words and not a transcript.

Context Discovery Analysis of UK General Election Debates
Scroll to the right to see the keywords etc.

What does this map represent about the debates: is it the number of occurences of the keywords, the relevance to the title.  So here is an opportunity for Henry or anyone else to tell us what it means.

About Andrew Wilcox

Andrew is an experienced user of MindManager who shares his knowledge and advice for free here. And provides commercial training and consulting on how to exploit MindManager and other mind mapping software applications in business, organisations and for individuals at Cabre For more information about Andrew please visit his Google + profile.

Tags: ,

2 Responses to “Applying Context Discovery to the Election Debate Mind Maps”

  1. Andrew Wilcox says:

    I have added 4 more Context Discoveries to the map.

    The sum of both debates.
    What did I hear: Gordon Brown, David Cameron and Nick Clegg, say.

    I am beginning to see how this tool works. Key words are identified (would like to know the criterai for this) and they are shown in their context. You also get a summary of the whole document.

  2. Hi Andrew,
    Here is a general comment on the use of keywords or rather key phrases. Key phrases are meant to serve multiple goals:

    1. They enable the reader to quickly determine whether the given article is in the reader’s fields of interest

    2. When key phrases are gathered in the cumulative index – the goal is indexing. They enable the reader to quickly find relevant content when the reader has a specific need.

    3. When a search engine uses key phrases the results are higher precision and smaller recall (number of documents). A search that matches a given query term in the keyword field will yield a smaller, higher quality list of hits than a search for the same term in the full text of the documents.

    Given the fact that we are inundated with information there is a need for tools that can automatically create key phrases. And paradoxically, although key phrases are very useful, only a small minority of the many documents provide key phrases.

    The principle for choosing key phrases is that the key phrase list is relatively short and it must contain only the most important, topical phrases for a given document.

    Our application, Context Organizer, treats a document as a set of phrases, which must be ranked in terms of relevancy. We view key phrase extraction as a classification problem.

    The task is to classify content in the document into one of two categories: either it is a key phrase or it is not a key phrase. We evaluate key phrases by the degree to which its classifications correspond to human-generated classifications.

    Our performance measure is precision (the number of matches divided by the number of machine-generated key phrases), using a variety of cut-offs for the number of machine-generated key phrases.

    A huge advantage of our approach is that the use of Context Organizer does not require costly training. The applied genetic algorithm and statistical methods are applicable across any types of documents. The technology is patented and in any benchmark tests performs accurately.

    Key phrases serve diverse goals the main one being the ability to capture the most important topics of the documents.

    And since a key phrase list can be read and judged in seconds the usefulness of this approach leads to practical applications. Once the key phrases are identified the contextual summaries are easily recognized and collected. The key phrases and summaries are a suitable candidates for mind mapping, generating briefings, reports, documents summaries, indexes, back-of-the-book lists for browsing, profiling user interests, personalization of subscriptions and many more applications.

    There are many ways to further customize summarization by creating keywords to watch and keywords to block lists. The biggest advantage of summarization is that it acts as content filtering applications. And in this sense it works in a similar way that we read – by scanning and skimming. Hope that this is somehow helpful.

Leave a Reply