The IBM Watson Tone Analyzer service is based on the theory of psycholinguistics, a field of research that explores the relationship between linguistic behavior and psychological theories. The service uses linguistic analysis and the correlation between the linguistic features of written text and emotional and language tones to develop scores for each of these tone dimensions.
Psycholinguists have worked to understand whether the words that we use in our day-to-day lives reflect who we are, how we feel, and how we think. After several decades of research in this area, it is now accepted in psychology, marketing, and other fields that language reflects more than just what we want to say. The frequency with which we use certain types of words can provide clues to our personality, thinking style, social connections, and emotional state.
Many people naturally read a message and judge the tones conveyed by the sender. But can a computer detect the tones disclosed by a message accurately and automatically? This is one of the many challenging questions to which researchers in the artificial intelligence and cognitive science fields are seeking answers. First with the Personality Insights service and now with the Tone Analyzer service, IBM is beginning to answer this question.
The discussion here explores the Tone Analyzer service on IBM Cloud under the Watson Services. I look at a few use cases where it is best applied and briefly discuss what it is, why was it introduced, and more importantly, how it is used.
Tone Analyzer service features
The Tone Analyzer service leverages cognitive linguistic analysis to identify a variety of tones at both the sentence and document level. You can use this insight to refine and improve communications. It detects three types of tones, including emotion (anger, disgust, fear, joy, and sadness), social propensities (openness, conscientiousness, extroversion, agreeableness, and emotional range), and language styles (analytical, confident, and tentative) from text.
Listed below are some of the features (or use cases) of the service:
- Conduct social listening: Analyze emotions and tones in what people write online, like tweets or reviews. Predict whether they are happy, sad, confident, and more.
- Enhance customer service: Monitor customer service and support conversations so you can respond to your customers appropriately and at scale. See if customers are satisfied or frustrated, and if agents are polite and sympathetic.
- Integrate with chat bots: Enable your chat-bot to detect customer tones so you can build dialog strategies to adjust the conversation accordingly.
Here’s a real-time illustration of it works:
Tone Analyzer Scorecard
The constituent parts of the three dimensions of Tone Analyzer Scorecard include:
- Emotional Tone: Many psychology models exist in literature to capture human emotions such as anger, fear, anticipation, surprise, joy, sadness, trust, and disgust. IBM Cloud has developed a model for inferring emotions from written text. Tone Analyzer captures the salient three among these that we found to be relevant for Tone analysis based on user studies. These include: cheerfulness, negative emotions, and anger. Cheerfulness refers to positive emotions like joy, optimism, contentment, inspiration, and happiness. Negative emotions include feelings of fear, disgust, despair, guilt, rejection, and humiliation. Anger is a type of negative feeling with strong intensity such as annoyance, hostility, aggression, hurt, frustration, and rage.
- Social Tone: Social tone includes aspects of social propensities in people’s personality. Tone Analyzer currently uses three social tones: openness, agreeableness, and conscientiousness, adopted from the Big Five personality model. Specifically, openness is the extent to which a person is open to experience a variety of activities, agreeableness is a tendency to be compassionate and cooperative towards others, and conscientiousness is a tendency to act in an organized or thoughtful way. We use these three dimensions to illustrate the openness, agreeableness, and conscientiousness of the writer as reflected in the text.
- Writing Style/Tone: Writing tone provides feedback on how analytical, confident, and tentative one’s writing is. Analytical tone shows a person’s reasoning and analytical attitude about things. Confidence tone indicates the degree of certainty exhibited by an individual toward something. Tentative tone shows the attitude of inhibition.
In addition, the Tone Analyzer service explains which words in the provided text contributed to which tone. Furthermore, it offers alternate word suggestions to refine the text to reflect desired tones. There is no minimum word requirement for the text input in order for the Tone Analyzer service to perform its functions well.
In conclusion, research has shown a strong and statistically significant correlation between word choice and personality, emotions, attitudes, intrinsic needs, values, and thought processes. Several researchers have found that people vary in how often they use certain categories of words when writing for blogs, essays, and tweets, and that these communication mediums can help predict different aspects of personality. Most of these prior works are based on finding psychologically meaningful word categories from word usage in writing. This research serves as the basis for IBM’s work on the Tone Analyzer service. Relying on the scientific findings from psycholinguistics research, IBM is working to infer people’s personality characteristics, their thinking and writing styles, their emotions, and their intrinsic needs and values from the words that they write. IBM uses its machine learning models to evaluate these characteristics by assessing various features of a person’s writing. This is highly achievable with the Tone Analyzer service on IBM Cloud.