Enriching semantic knowledge bases for opinion mining in big data applications.

Area of Science:

• Natural Language Processing
• Artificial Intelligence
• Web Intelligence

Background:

• Opinion mining relies on semantic knowledge bases, which often lack contextual depth.
• Existing sentiment lexicons and affective resources may not capture nuanced opinions effectively.
• High-throughput big data applications require robust and context-aware sentiment analysis.

Purpose of the Study:

• To present a novel method for contextualizing and enriching large semantic knowledge bases for opinion mining.
• To enhance the performance of sentiment analysis in Web intelligence and big data platforms.
• To improve the applicability of affective resources like SenticNet.

Main Methods:

• Identifying ambiguous sentiment terms.
• Extracting context information from domain-specific training corpora.
• Grounding contextual information to structured knowledge sources (ConceptNet, WordNet).

Main Results:

• Significant improvement in polarity classification accuracy using an enriched SenticNet.
• Quantitative evaluation demonstrated the effectiveness of the enrichment process.
• Qualitative evaluation provided insights into concept grounding and enrichment quality.

Conclusions:

• The proposed method effectively contextualizes and enriches semantic knowledge bases for opinion mining.
• Enriched affective resources lead to enhanced sentiment analysis performance.
• The approach offers a valuable contribution to opinion mining in big data environments.