This means that when calculating whether a particular statement in social media is positive or negative, the use of these words is used by these sentiment analysis tools as an indication that the statement is negative towards the topic of the statement - such as a topic, issue, individual or organisation.
I've checked a number of sentiment dictionaries online and found that both 'disabled' and 'disability' appear frequently as negative terms. However I have not yet been able to confirm whether any sentiment analysis products treat these words in this manner.
This disturbs me, given the efforts of governments and civic organisations in Australia and many other countries to remove negative stigma attached to the word 'disabled', even given its potential application in statements such as 'their system has been disabled'.
It also concerns me that agencies engaging online about disabilities or with disabled people, might accept that the sentiment reported by their social media monitoring tools indicates negativity where in actuality no negativity exists.
I would caution government agencies using automated sentiment analysis tools to get to know they work and check how terms such as 'disability' and 'disabled' are treated in these systems.
I'd welcome comments from makers of sentiment analysis tools to confirm how they treat these words or from agencies using automated sentiment tracking if they've seen these words or others rated negatively or positively in ways which might be misleading and misrepresent the actual sentiment.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.