Cooperation and service offer
We offer services in research, design and implementation of methods for automatic processing of text data in the form of natural language. Most of the methods are working on inflected languages (Czech, Slovak and others) and English. The methods are designed to process large volumes of the data and are based on machine learning, which shows a good ability to adapt to unfamiliar or new phenomena in the texts.
We focus mainly on the following tasks:
- Language Identification (27 languages) – suitable even for very short texts such as blog posts or Internet discussions.
- Reconstruction of the Text – includes adding diacritics and punctuation, corrections of typographical errors and non-literary forms.
- Named Entity Recognition – allows to find words or phrases with important meaning (e.g. persons, cities, states, product names, companies, date and time values, etc.).
- Sentiment Analysis – consists in automatic identification of the sentiment of sentences or short texts. We can recognize for instance either the categories (positive, negative or neutral) or assign a value from 1 to 10.
- Intelligent Search – facilitates searching in texts using additional information (e.g. stemming, semantics, named entities, etc.).
- Results Analysis – helps to analyze the search results, track trends and correlation of events in the results.
- Automatic Document Classification – is automatic identification of document categories according to the content (e.g. politics, sports, weather, etc.). We are able to do multi-label classification, i.e. to classify documents into multiple classes, which is beneficial for commercial usage.