CIDOBOT
This proof of concept (PdC) implements a system of artificial intelligence to automate the themed classification of ads at bulletins and official newspapers. The system, integrated at the seeker CIDO, utilises neuronal meshes trained with historic data to improve the efficiency and the accuracy at the documentary management.
Singled out results:
- Treatment automated of 700 daily ads
- Annual saving amounting at a month of human work
- Tall rank of accuracy at the classification
- Architecture escalable and replicable
Recommendation: Cement the system CIDOBOT and explore new functionalities like the generation of digests, the classification for ODS and the documentary management automated.
Promotor: Barcelona's deputation
Status: Finalitzat
veure-documentacio