Hello, and thanks for checking out our blog. We’re planning on keeping the blog pretty current, and posting several times per week. We’ll be discussing the rapid evolution of Social Media and where we fit into that world, as well as some detailed discussions around the uses of text analytics and some of the up and coming technology changes around it. For our initial post, I’m going to focus on that feature of our system that more or less made Lexalytics – Sentiment Analysis.
Sentiment Analysis is one of the hot features in text processing these days, but it often falls into the “cool” rather than required list of features. This is mostly because people aren’t really sure how to use it to augment their business offerings, so let’s use this post to take an in depth look at one really interesting application of sentiment analysis that can affect the bottom line. Most of us have spent time on review sites like epinions and are pretty comfortable and happy with the user ratings for the products on these sites. The next step in these product reviews would be to introduce a similar quantitative score for the features of the products that we’re researching. Applying sentiment scoring to the feature or attribute level allows for this more detailed review of a product. Let’s take the example of a digital camera. Various consumers have different hot button features; one might care more about the battery life of the camera, while someone else might care more about the zoom. By mining reviews for these attributes and then scoring the reviews at this atomic level we’re able to provide more detail than a product level score of 4.5 stars. Why site owners should care is that this will increase the volume and quality of information on their sites, improve the stickiness and in the end increase ad revenue. Sentiment is still a relatively new technology in the text space, so the uses of it to affect the bottom line are just now being recognized.