Analytics service examples

Example of E-health

SOLUTION: Fast and precise search

PROBLEM

The development of our software started out in Software Technology and Applications Competence Centre (www.stacc.ee) in order to analyse digital health data. Despite the big amounts of data and exciting scientific perspectives, digital health stories are dominantly free textual that makes their analysis very complicated. To carry out scientific work more effectively, one needed an environment where different hypotheses could be quickly tested along with the familiarisation with the datasets.

SOLUTION

Use of TEXTA Toolkit helps STACC scientists and data analysts to surf in the millions of health documents, to search and identify in milliseconds the interesting cases, analyse the work of the developed algorithms etc.

Classification of comments in newspaper Õhtuleht

SPECIAL SOLUTION: systematisation of current information

PROBLEM

There is a big number of comments in the commentary, which are being pre-filtered and could split the commentary, veering off the topic, going personal or even bringing a lawsuit against Õhtuleht.

SOLUTION

The task was made complicated by the richness of language, nonsubjection of comments to the grammar and changing of language in time. We took it all into account and created a software solution, which automatically analyses the comment and issues the tags describing the latter. The solution can detect in an instant, whether the comment contained something insulting, threatening, obscene or racist.

Example of Inforegister

SPECIAL SOLUTION: Systematisation of current information

PROBLEM

Representatives of different realms need information on the judicial reliability of the counterparty. At present, one could check all the public judicial decisions in the registry of judicial decisions on the website of the Ministry of Justice. However, this enables only manually (working case by case) to verify judicial background of the enterprises and entrepreneurs, which means that to collect information one has to read through all the judicial decisions concerning the enterprise.

SOLUTION

Output of the project is an analysis engine of judicial decisions, which regularly processes the documents of the registry of judicial decisions and identifies information describing the result of lawsuits, which will be then readily made available for the customer. To solve the task, we used partitioning of information to find relevant information from the text – who were the parties to the lawsuit and how did the legal process end.

Ready-made solutions examples

Classification of comments in newspaper Õhtuleht

SPECIAL SOLUTION: systematisation of current information

PROBLEM

There is a big number of comments in the commentary, which are being pre-filtered and could split the commentary, veering off the topic, going personal or even bringing a lawsuit against Õhtuleht.

SOLUTION

The task was made complicated by the richness of language, nonsubjection of comments to the grammar and changing of language in time. We took it all into account and created a software solution, which automatically analyses the comment and issues the tags describing the latter. The solution can detect in an instant, whether the comment contained something insulting, threatening, obscene or racist.