Table 2 Summary table of anomaly detection techniques with the traditional approach
From: Intelligent systems for sitting posture monitoring and anomaly detection: an overview
Techniques | Advantages | Limitations | Refs. |
|---|---|---|---|
Rule-based | Application simplicity High inference speed Low computational cost Transparent and explainable process | Expert knowledge required Difficulty in capturing complex relationships Limited to a low number of postures and differentiated from each other | |
Statistical | Interpretability of results They provide probability estimations for each class Reasonable training time | Assumption that the data follow a probability distribution Expert knowledge required Sensitive to irrelevant features Limited performance with complex data | |
Intelligent | Ability to capture complex relationships Greater accuracy Adaptable to input changes Ability to ignore redundant features | Need for large data sets Increased training time and computational resources Difficulty in selecting hyperparameters Lack of interpretability of the decision making process | [5, 15, 17, 19, 24,25,26,27,28,29,30,31,32, 66, 68, 72, 75,76,77,78,79,80,81,82,83,84,85] |