Abstract:Happiness computing based on large-scale online web data and machine learning methods is an emerging research topic that underpins a range of issues, from personal growth to social stability. Many advanced Machine Learning (ML) models with explanations are used to compute the happiness online assessment while maintaining high accuracy of results. However, domain knowledge constraints, such as the primary and secondary relations of happiness factors, are absent from these models, which limits the association between computing results and the right reasons for why they occurred. This article attempts to provide new insights into the explanation consistency from an empirical study perspective. Then we study how to represent and introduce domain knowledge constraints to make ML models more trustworthy. We achieve this through: (1) proving that multiple prediction models with additive factor attributions will have the desirable property of primary and secondary relations consistency, and (2) showing that factor relations with quantity can be represented as an importance distribution for encoding domain knowledge. Factor explanation difference is penalized by the Kullback-Leibler divergence-based loss among computing models. Experimental results using two online web datasets show that domain knowledge of stable factor relations exists. Using this knowledge not only improves happiness computing accuracy but also reveals more significative happiness factors for assisting decisions well.

What problem does this paper attempt to address?

The paper attempts to address the issue that existing happiness computation models, while maintaining high accuracy, lack domain knowledge constraints regarding the primary and secondary relationships of happiness factors. This leads to a weak correlation between the computation results and the actual reasons. Specifically, the paper focuses on the following aspects: 1. **Lack of domain knowledge constraints**: Existing machine learning models, although capable of achieving high prediction accuracy in online happiness assessment, lack domain knowledge constraints regarding the primary and secondary relationships of happiness factors, resulting in low consistency and credibility of explanations. 2. **Insufficient explanation consistency**: Different models provide significantly different explanations for happiness factors, making it difficult to determine which model's explanation is correct, thereby affecting the accuracy of decision-making. 3. **Lack of guidance in model training**: Existing model training mainly relies on labeled data and lacks constraints on model explanations, leading to situations where models may make correct predictions for the wrong reasons, thus undermining the effectiveness of decisions. To address these issues, the paper proposes the following objectives: - **Introduce domain knowledge**: By introducing the primary and secondary relationships of happiness factors as domain knowledge, improve the consistency and credibility of model explanations. - **Optimize model training**: Utilize domain knowledge to constrain model training, ensuring that the model not only predicts accurately but also makes predictions for the right reasons. - **Validate the effectiveness of the method**: Through experiments, validate the effectiveness of the proposed method in improving the accuracy and consistency of happiness computation. In summary, the paper aims to improve existing happiness computation models by introducing the primary and secondary relationships of happiness factors as domain knowledge, thereby enhancing the consistency and credibility of explanations while maintaining high accuracy, to better support decision-making.

Primary and Secondary Factor Consistency as Domain Knowledge to Guide Happiness Computing in Online Assessment

Quantitatively Interpreting Residents Happiness Prediction by Considering Factor–Factor Interactions

A hierarchy-based machine learning model for happiness prediction

Developing a machine learning‐based instrument for subjective well‐being assessment on Weibo and its psychological significance: An evaluative and interpretive research

Predicting life satisfaction using machine learning and explainable AI

Prediction of Adolescent Subjective Well-Being: A Machine Learning Approach

Identifying the top predictors of student well-being across cultures using machine learning and conventional statistics

Sensing Psychological Well-being Using Social Media Language: Prediction Model Development Study

Human Wellbeing and Machine Learning

Evaluating Machine Learning Stability in Predicting Depression and Anxiety Amidst Subjective Response Errors

Construction And Measurement Of Artificial Psychological Emotion Model For Consumers' Online Shopping Behaviors

Identifying the most crucial factors associated with depression based on interpretable machine learning: a case study from CHARLS

Automatic Expansion of Domain-Specific Affective Models for Web Intelligence Applications

Improving Human Happiness Analysis Based on Transfer Learning: Algorithm Development and Validation

High School Emotional Well-Being: Identifying Factors And Creating A Working Model For Prediction

The Effect of Individual-Level Factors and Task Features on Interface Design for Rule-Verification Crowdsourcing Tasks

Stacking Model for Optimizing Subjective Well-Being Predictions Based on the CGSS Database

UNIFYING ALGORITHMIC AND THEORETICAL PERSPECTIVES: EMOTIONS IN ONLINE REVIEWS AND SALES

Fostering Knowledge Exchange in Digital Communities: Psychological Determinants of Sharing in Q&A Platforms

WellDunn: On the Robustness and Explainability of Language Models and Large Language Models in Identifying Wellness Dimensions

Identifying discernible indications of psychological well-being using ML: explainable AI in reddit social media interactions