Technology leaves no stone unturned. When the human brain and technology join hands then the outcome is impeccable. In a similar phase, the University of California San Diego School of Medicine has come up with the idea of detecting loneliness. Researchers have used Artificial Intelligence algorithms to detect the solitude and quantify them accordingly.
Over the past two decades, there prevails the trend of loneliness among the population. Solitude has given rise to many health-related issues. It is clearly evident from the drug case, suicide rates, depression rates, and other health-related problems. Loneliness seems to be vulnerable to senior citizens. The study conducted by UC San Diego implies that senior citizens tend to have loneliness rates rising to 85%. This problem can be resolved properly when the depth and breadth of the issue are identified.
Data collection would be quite challenging because of biased responses and social stigmas among the respondents. To overcome this limitation, a better metric is designed to quantify loneliness with the help of Natural Language Processing (NLP) and Machine Learning. This method is believed to receive genuine responses from people. Ellen Lee, Assistant Professor of Psychiatry at UC San Diego, along with his research team have conducted a study to detect loneliness. 80 participants have been selected between the ages of 66 to 94. In line with the suggestion of a Ph.D. Varsha Badal, the former author of the study, long answers are taken during the interviewing process. Then with the help of Natural Language Learning and Machine Learning, the indicator of loneliness such as speech patterns and choice of words are detected. NLP and Machine learning, analyze those answers meticulously and identifies solitude even with subtle features like emotions. On the other hand, when a similar emotional analysis is done by humans, there exist possibilities for error due to improper training, biasing, and lack of consistency.
According to the study, notable differences are clearly evident among lonely and non-lonely respondents. The responses are long and express desolation among the lonely respondents. Findings show there exists an unlikeness or hindrance in admitting solitude between male and female respondents. Men were less likely to acknowledge loneliness and more likely to use words related to Happiness or terror. The result of the study implies that loneliness can be detected with the help of Speech pattern analysis and can be treated in earlier stages when diagnosed properly. Machine learning proves to be 94% accurate in determining the solitude among people. Further research can be focused on finding its reliability and robustness to make replicate this model for further findings.
