Social sensing applications and case study using acoustic Arabic opinion mining

Recently social sensing has gained a growing attention with many promising applications in smart cities. This paper reviews application domains of social sensing to provide context-aware services in smart cities including healthcare, wellness and safety, transportation, business, and environmental monitoring. Moreover, a case study is presented to understand human opinions expressed in Arabic spontaneous speech. Thirty four features are extracted to discriminate each opinion from audio signals and several machine-learning classifiers are applied to detect its polarity. To deal with a real-world problem, the dataset is composed of audio clips with diverse characteristics including speakers’ nationalities and ages, environments and topics.