Name
Abstract | ||
|---|---|---|
The surge and value of unstructured text is attracting substantial research and industry attention. Subsequently we are witnessing novel techniques and algorithms that are performing increasingly sophisticated text mining tasks. However the majority of such techniques are opaque, making it hard to trace the provenance of the analytical task on hand. We propose Catalyst, a framework to automatically transform, enrich and expose text into a linked graph-based layer to enable more transparent processing and access to the text elements. In brief, Catalyst extracts text dependencies, performs sentiment analysis, detects semantic relatedness, and links the text elements into a semantic triple-store that enables an easy access to the text entities through direct query functionalities. We plan to evaluate the performance of Catalyst by processing a dataset of user reviews around the dimensions of an evaluation model deployed in the context of e-government services. |
Year | Venue | Keywords |
|---|---|---|
2017 | AMCIS 2017 PROCEEDINGS | Text analytics, user feedback, decision support, performance management, linked data, semantic web |
Field | DocType | Citations |
Data science,World Wide Web,Text mining,Computer science | Conference | 0 |
PageRank | References | Authors |
0.34 | 0 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Fouad Zablith | 1 | 236 | 15.15 |
| Bijan Azad | 2 | 204 | 8.50 |
| Ibrahim H. Osman | 3 | 815 | 94.23 |