Machine Learning Classification of Adult Attachment Styles Based on Dyadic Behavioral and Emotional Indicators
Bibliographic record
Abstract
The objective of this study was to develop and evaluate machine learning models capable of classifying adult attachment styles using multimodal dyadic behavioral, emotional, and physiological indicators derived from real-time couple interactions. A cross-sectional observational design was implemented with adult romantic couples recruited from community settings in Canada. Both partners in each dyad participated in standardized interaction tasks designed to elicit attachment-relevant behaviors, including conflict discussion and support-seeking exchanges. Adult attachment styles were assessed using validated self-report measures and used as supervised learning labels. Multimodal data were collected, including behavioral coding of dyadic interactions, self-reported emotional responses, physiological indices of autonomic regulation, and paralinguistic and facial-expression features extracted from audio–video recordings. Machine learning pipelines incorporated data preprocessing, feature extraction at the dyadic level, dimensionality reduction, and model training using multiple classification algorithms. Stratified dyad-level cross-validation and hyperparameter optimization were applied to ensure robust generalization and prevent data leakage. Non-linear and ensemble-based models significantly outperformed linear classifiers in attachment style prediction, with neural network and gradient boosting models achieving the highest accuracy and area under the receiver operating characteristic curve. Dyadic emotional synchrony and observed behavioral responsiveness emerged as the strongest predictors of attachment style classification, followed by self-reported attachment dimensions. Physiological and paralinguistic indicators provided incremental predictive value when integrated with behavioral features. Cross-validation analyses demonstrated high stability across folds, and misclassification patterns primarily occurred between theoretically adjacent attachment styles, indicating construct-consistent overlap rather than random error. The findings demonstrate that adult attachment styles can be accurately classified using machine learning models trained on multimodal dyadic interaction data, supporting a relational and interaction-based conceptualization of attachment. This approach offers theoretical advances in attachment research and practical implications for objective assessment and intervention planning in couple and relational contexts.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.002 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".