Use CasesUse Cases
Find the current sentiment and add a positive spin to new Blog Posts.
- Social Media Sentiment
Track how your company is perceived and take action.
- Product Reviews
Correct or expand the reaction to your product.
Automated Sentiment Analysis
We have an extensive compiled entity dictionary as part of our knowledge base, and we update it daily. Our knowledge base automates the NLP (Natural Language Processing) sentiment analysis for overall text, entities, and keywords. By selecting a domain area such as news, product reviews, or social media, the service tailors to the domain and refines the sentiment accuracy. We have been refining our algorithms to reflect a wide spectrum of polarities. Natural Language Understanding (NLU) allows us to continually train our service to recognize the similarities within the text in the domain and become more accurate with the results.
Several customers have special entities and keywords particular to their business. We have modified our automated text analysis to allow for user-specified customer dictionaries for additional entities. When selected, the custom entity dictionary is used to identify and extract entities that may not yet be considered as entities in the current dictionary. You can update the dictionary as these new entities become part of your business.
Smart Sentiment Service assigns an overall sentiment: positive, negative, neutral, or no sentiment to the original text. The sentiment is based on a weighted comparison of all sentiment-words in the initial text document. Sentiments are based on a weighted comparison of all natural language words denoting or evoking a sentiment, such as “good”, “bad”, “lovely”, “splendid”.
The “neutral” and “no sentiment” are not the same. “No sentiment” means there is not enough sentiment words to determine a sentiment. A “neutral” sentiment results from the presence of both “positive” and “negative” sentiments that neutralize each other.
Smart Sentiment Service extracts entities (persons, places, or organizations) from the text, and assigns a sentiment per entity. The sentiment uses the local context or sentiment-words in the vicinity of the entity to help determine the sentiment score. We identify the prime category for the entity and provide additional categories. The additional categories are possibilities for new audiences or business opportunities.
The keyword sentiment analysis captures your keywords. We assign a sentiment to the keywords you enter. Keywords can be any phrase or word and are not limited to person, place, or thing (entities). The keyword score shows the importance of the keyword to the original text. Having an individual keyword sentiment score, and an entity sentiment score, and overall score gives a more complete picture of the sentiment around your topics and categories.