Amazon’s new AI Textract is one of the hottest topics today. The program can process millions of pages in just a few hours. This is a huge breakthrough in the text analytic field. Just imagine how much easier it is going to be for a small or a medium company to process big amount of information without losing hundreds of human hours. The development of AI text analysis is paramount as it enables us to improve the process of decision making and cost saving. For example, if you have a tool that analyses unstructured text information faster and reduces the cost for the person needed to do that job and also you can focus on tasks that require creativity and leave the time-consuming details to automation. The extracted text and data can easily be used to build smart searches on large archives of documents or can be loaded into a database for use by applications such as accounting, auditing, and compliance software.
Text data analytics has always been an important part of business development. That’s why there are so many investments in this kind of technology. Imagine what can be achieved with better text analyze. For example, HyperAspect vast experience shows how text analyzing can help in medicine and mental therapy. Several UK based hospitals decided to track their patients’ condition by making them fill a digital diary. They have ended up with a vast set of unstructured text, where patients were describing how they feel on a daily basis. In order to process all of the date, they had to employ a separate person to go through all diary entries, determine which entries are important and pass them forward to medical professionals. Our customer was looking for a solution that will reduce the cost of extra labor hours.
Thanks to the HyperAspect machine learning platform – HyperLearn, we were able to provide sentiment scoring for the specific treatment that was applied, once that’s done we have funneled the data to our text exploration workbench TeX where we derived a trend line that unambiguously shows how a patient feels about his progress from the treatment applied. On top of that, we were able to correlate expression that potentially can impact the treatment. A patient treated for sleeping disorder showed significant trend line spike, but our system was able to point that the spike occurred right after the patient stated that she purchased a brand new bed. This data was transferred to our client system as false-positive index alongside the trendline for further processing by medical professionals. After the deployment of our solution the customer was able to abandon any type of manual pre-processing, the medical professionals initially was using their existing system in order to read the extracted data. Soon after the final deployment, the customer developed an alerting system that contacts medical professional once specific sentiment trendline threshold is reached.
In the future, the secrets that are hidden in a big amount of text data will be even more important for the companies. Technology could one day be used to help people who can’t speak or to prevent psychological problems analyzing unstructured text data. Big companies like Amazon and Google working hard to develop and update text analytics for more effective work.
Do you find this article useful? Write to us for more interesting case-studies from our work.