SPEEDUP BRIDGE ALFA
Sentimenti was awarded a grant in accordance with the project implemented by SpeedUp Bridge Alpha limited liability company under:
Activities
- Intelligent Development Operational Program 2014-2020,
- Measure 1.3: R&D financed with the participation of capital funds
- Sub-measure 1.3.1: Support for R&D Projects in the preseed phase by proof of concept funds – BRIdge Alpha.
Project
“Sentimenti prediction of changes in the rates of financial instruments based on emotion contained in the text”.
Purpose of the project
The purpose of an R&D Project is a set of organized actions of a scientific or research and development nature to achieve the goal of the Grant Project, with commercial potential both for those involved in its implementation and for investors interested in commercializing its results, carried out in the PoP or PoC phase.
Value of the project
- 1 000 000 PLN, in particular:
- NCBiR: 800 000 PLN
- Fundusz SpeedUp Bridge Alfa: 200 000 PLN
“SENTIMENTI” – emotion analyzer in the written word
Project number: POIR.01.01.01-00-0472/16
Start Date: 01.06.2017, End Date: 31.10.2019
Implemented under: Operational Program Intelligent Development 2014-2020, co-financed by the European Union from the European Regional Development Fund.
Measures:
Priority axis 1 – Support for conducting R&D works by enterprises, Action 1.1 – R&D projects of enterprises, Sub-action 1.1.1 – Industrial research and development works performed by enterprises.
Main objective of the project
To develop SENTIMENTI software to analyze written content, particularly online content, for emotion and emotional overtones.
About the project
The software, as the first in the Polish language, will be designed based on one of the theoretical models, assuming the existence of several distinct culturally universal and biologically determined emotions, namely. EKMAN’S MODEL (linking emotions to facial expression), PLUTCHIK’S EMOTION MODEL (allowing to capture both the intensity of emotions and the degree of their similarity) or GENEVA’S EMOTION WHEEL MODEL (assuming an approach of emotional dimensions by placing individual groups of emotions within emotion sign and control and having a five-level scale of emotion intensity).
Basing the technology on the indicated models will allow characterizing utterances in terms of evaluating polarity of overtones and basic emotions – unlike current solutions, which offer to exclude polarity of overtones as negative, positive, or neutral. Descriptions of this type have broad applications in media monitoring, analytics, marketing and instant messaging. In addition, the new technology allows integration with systems of users interested in content analysis (e.g., social media). The solution will enable processing of content written on portals, chats, forums performing analysis that recognizes the emotions expressed in them and the polarity of overtones, in the form of reports of emotional intensity of the studied content.
The algorithms implemented in the software will consider the context of the occurrence of the words and phrases used, based on extensive, efficient local analysis grammars based on natural language engineering methods. The software allows easy, semi-automatic tuning by machine learning methods to a selected set of subject domains for maximum quality of action.
Project financing
Project value: PLN 5,843,511.50, Grant value: PLN 4,427,520.09