When is a dog a dog? When is cool cool?

Here at Feedback Ferret, we know only too well that training a machine in Natural Language Processing is complex.

Thinking about the word “dog”, the machine is given various characteristics that help it detect the fact that a dog is being talked about. For example, ‘barking’, ‘paws’, ‘4 legged’, ‘furry friend’. Thinking about the word “cool”, the characteristics of the word may be ‘not warm’, ‘cold’, ‘temperature’.

But what does the machine do when someone comments “this car’s a dog”?

dog of a car

Or “the hotel room was really cool”?

Cool hotel room

How does a machine, that has been taught automated rules, understand that in fact the car quality is poor and the hotel room may in fact be rather awesome?

This is when humans need to intervene in the automated process and, through rigorous quality control and ongoing updates to the coding of phrases, they can develop and evolve the machine to understand and correctly code ambiguous phrases.

Back in 2002, Feedback Ferret created a text analytics solution which understands the words that customers use to express their expectations and experience in a way that had never been possible before. The accuracy of our unique approach to text analytics speak volumes for the “machine + human” approach over 100% automated solutions based on Natural Language Processing. We regularly see very high accuracy rates and clients are often astounded by how our text analytics engine understands and correctly codes complex phrases that appear in their customer feedback.

So, if you’re thinking about using text analytics to understand your Voice of Customer comments, make sure you choose a solution that can distinguish between a dog & a dog, and cool & cool.

Or why not just contact us?