myPDA - my Policy Discourse Analyser

Step 1 - Brief introduction

This web page was built to perform automated discourse analysis of neighbourhood policy text. It aims to provide a score indicative of level of commitment to themes characterised by trigger phrases. A corpus of trigger phrases was extracted from European Neighbourhood Policy texts and scores were assigned by a panel of 3 independent curators with different backgrounds. These default trigger phrases can be overwritten with a user defined set.

This software is for research use only and is provided "AS IS". No claims towards the validity of its outcome are made by the authors and no liability arises from making this software available to you.

This web page takes your input: (1) a piece of input text on which discourse analysis will be performed, (2) a pre-defined list of "stop-words" (explained below) that you can modify or replace, and (3) a trigger-and-score list, representing the corpus of trigger phrases introduced above. After collecting these inputs, the underlying algorithm will display analysis results and generate a score histogram and some statistics. The final step allows manual curation of the outcome to correct for any mistakes the algorithm has made.

Alternative writings

The software will try and deal in an intelligent way with how words and phrases are written. For instance, it won't matter if the text (1) or the trigger-and-score list (3) uses alternative writings, hyphens, or cases. For example, this means the following forms will all be recognised as 'cooperation' and be awarded the corresponding score:

	 co-operation
	 cooperation
	 Co-Operation

Hence, the following two sentence parts will be scored identically:

	 ...relationship, going beyond co-operation, to involve a significant...
	 ...relationship, going beyond cooperations, to involve a significant...

Stop-words

The stop word list (defaults apply, but can be overwritten) is the list that contains words that will be explicitly ignored by the algorithm in both the input text (1) and the trigger-and-score list (3). This is convenient in the sense that the following forms will both be scored according to trigger 'to go beyond cooperation':

	 ...relationship, going beyond a co-operation, to involve a significant...
	 ...relationship, going beyond co-operation, to involve a significant...

tenses and inflections

Finally, the algorithm makes efforts to recognise inflections and tenses. This means the following forms will be recognised as 'cooperation':

	 a co-operation
	 cooperations
	 to cooperate
	 we cooperated
	 cooperating

Hence, the following two sentences will both trigger on 'to go beyond cooperation':

	 ...relationship, to go beyond co-operating, to involve a significant...
	 ...relationship, going beyond co-operation, to involve a significant...

How to use this tool

This application is built as a wizard that takes you through the several steps of input and analysis. To get started, click the 'Next' button at the bottom or at the top of the page.


Step 1 - Brief introduction This web page was built to perform automated discourse analysis of neighbourhood policy text. It aims to provide a score indicative of level of commitment to themes characterised by trigger phrases. A corpus of trigger phrases was extracted from European Neighbourhood Policy texts and scores were assigned by a panel of 3 independent curators with different backgrounds. These default trigger phrases can be overwritten with a user defined set. This software is for research use only and is provided "AS IS". No claims towards the validity of its outcome are made by the authors and no liability arises from making this software available to you. This web page takes your input: (1) a piece of input text on which discourse analysis will be performed, (2) a pre-defined list of "stop-words" (explained below) that you can modify or replace, and (3) a trigger-and-score list, representing the corpus of trigger phrases introduced above. After collecting these inputs, the underlying algorithm will display analysis results and generate a score histogram and some statistics. The final step allows manual curation of the outcome to correct for any mistakes the algorithm has made. Alternative writings The software will try and deal in an intelligent way with how words and phrases are written. For instance, it won't matter if the text (1) or the trigger-and-score list (3) uses alternative writings, hyphens, or cases. For example, this means the following forms will all be recognised as 'cooperation' and be awarded the corresponding score: co-operation cooperation Co-Operation Hence, the following two sentence parts will be scored identically: ...relationship, going beyond co-operation, to involve a significant... ...relationship, going beyond cooperations, to involve a significant... Stop-words The stop word list (defaults apply, but can be overwritten) is the list that contains words that will be explicitly ignored by the algorithm in both the input text (1) and the trigger-and-score list (3). This is convenient in the sense that the following forms will both be scored according to trigger 'to go beyond cooperation': ...relationship, going beyond a co-operation, to involve a significant... ...relationship, going beyond co-operation, to involve a significant... tenses and inflections Finally, the algorithm makes efforts to recognise inflections and tenses. This means the following forms will be recognised as 'cooperation': a co-operation cooperations to cooperate we cooperated cooperating Hence, the following two sentences will both trigger on 'to go beyond cooperation': ...relationship, to go beyond co-operating, to involve a significant... ...relationship, going beyond co-operation, to involve a significant... How to use this tool This application is built as a wizard that takes you through the several steps of input and analysis. To get started, click the 'Next' button at the bottom or at the top of the page.