T-LAB Plus 2016

T-LAB Plus 2015
14 September 2015
T-LAB Plus 2017
20 January 2017

T-LAB Plus 2016

T-LAB Plus 2016 was released on April 22nd 2016.

Listed below are some of the key improvements and new features:

1 - Now eleven different file formats - including PDF documents - can be processed either as a single file or as a collection of documents.



- The image-only PDF files must be processed using OCR software first;

- HTML files can be imported by using the Corpus Builder module only.

2 - Two user profiles are now available: beginner and expert. When the first (i.e. beginner) is selected, the user is allowed to perform any analysis without being asked to choose between the advanced options.


3 - Whenever analysing word co-occurrences and/or exploring clustering solutions, a new tool named Graph Maker allows the user to easily create and export several new dynamic charts and graphs, some of which are built with the D3 library.


4 - Every time a tool for exploring similarities and differences between corpus subsets or between thematic clusters is used a new button is available which allows the user to view a preview by means of a dynamic tree map.


5 - An additional algorithm for the thematic analysis (i.e. unsupervised clustering) of text segments and documents is now available which complements the bisecting Kmeans algorithm implemented in T-LAB more than ten years ago.
The new algorithm uses the PDDP (i.e. Principal Direction Divisive Partitioning) method proposed by Daniel Boley (1998) to initialize a k-means clustering procedure.

This way the user of T-LAB with an expert profile is able to compare different solutions for the same clustering problem, e.g. compare the quality of clusters obtained by two different algorithms applied to the same data tables.



Fig. 1