AI technologies used in opioid research by all included studies (n=81), based on Brownlee’s Tour of Machine Learning Algorithms26
| AI technology | Number of studies* |
| Ensemble algorithms | 48 |
| Random forest | 29 |
| Gradient boosting machine | 7 |
| Stochastic gradient boosting | 5 |
| eXtreme Gradient Boosting | 4 |
| Boosted tree | 1 |
| Adaboost | 1 |
| Ensemble (unspecified) | 1 |
| Natural language processing | 37 |
| Natural language processing | 29 |
| Natural language processing and the Biterm Topic Model | 5 |
| Natural language processing and machine learning | 2 |
| Natural language processing and topic modelling (Latent Dirichlet Allocation) | 1 |
| Deep learning algorithms | 23 |
| Neural network | 10 |
| Deep neural network | 5 |
| Convolution neural network | 4 |
| Recurrent neural network | 4 |
| Instance-based algorithms | 22 |
| Support vector machine | 18 |
| K-nearest neighbour | 4 |
| Regularisation algorithms | 16 |
| Elastic Net | 10 |
| Least Absolute Shrinkage and Selection Operator | 6 |
| Regression algorithms | 13 |
| Logistic regression | 10 |
| Regularised regression | 1 |
| Multinomial regression | 1 |
| Boosted regression | 1 |
| Bayesian algorithms | 5 |
| Naïve Bayes | 4 |
| Bayesian belief network | 1 |
| Decision tree algorithms | 4 |
| Decision tree | 4 |
| Clustering algorithms | 1 |
| Bi-k-means clustering modelling | 1 |
| Other | 1 |
| Hybrid combining multiobjective optimisation and support vector regression | 1 |
*Some studies reported the use of multiple types of AI technologies in a single study; thus, the total number is greater than 81.
AI, artificial intelligence.