Challenges and how Annova overcame them.

Annova’s task consisted of labeling the sentiment of a consumer towards a given tag while consuming a product, and providing reasoning when necessary. A very good knowledge of English was necessary for this project 50,000 plus entries were processed with
99.99% accuracy.

The key challenge was the enormous amount of unstructured data to be traced online and label the it for further consumption of the client without any bias and errors. We had to read out all the comments – positive, neutral, negative and look at the complexity of the sentiment in the same comment. For example, if there is a comment like, ” the burger was yum but the mayo seemed too tangy”, segregation of the sentiment is critical and it’s a complex task for the model. We had our model trained to judge the mood of the people in order to come up with a holistic analysis so that there are no room for any bias.

We completed this complex project on time, the accolade was more business from the same client. This kind of assignments give Annova a confidence boost, make Annova more ready for any kind of challenge while we sharpen our technology tools we work with.