| | Technique | Findings |
| | KNN | The feature similarity is generated by KNN algorithm application used to predict the values of any new data points, assigning a value that closely resembles the points in the training set. Thus, for low and damp pressure, animal behaviour is overactive, temperature is cold, falling of leaves is high, time is early morning, and depth can be different, having a magnitude between 4.1 and 5.14 shows a high possibility of occurrence of earthquake (refer to Figure 3). |
| | SVM | Once the mapping is done by SVM with regression using linear magnitude and depth with cross-validation, although the time monitored is different, i.e., time is both morning and evening, for low pressure, animal behaviour is overactive, temperature is very cold, falling of leaves shows high, but the variation in depth is negligible, location is negligible, and falling of leaves having magnitude between 4.1 and 4.9 shows high possibility of occurrences of earthquake (refer to Figure 4). |
| | XGBoost | Applying XGBoost gradient on preprocessor data generates relationship and correlation on attributes, and values show that even though the atmospheric pressure is damp and dry, temperature is very cold, time is morning, the falling of leaves is medium and high, and both dry and damp pressures show cold or very cold temperature within a range of closer depth and location having a magnitude between 4.1 and 5.1 which shows high possibility of occurrences of earthquake (refer to Figure 5). |
| | Decision tree | The findings of the graphs by decision tree are constructed suggesting that there is a strong relationship on possibility of occurrences of earthquake having the ranges of magnitude ranging from 4.0 to 5.04, animal behaviour is more active than normal, atmospheric pressure is dry, temperature is either very cold or cooler than normal, falling of leaves is high, and depth variation is minimum, showing high possibility of occurrences of earthquake (refer to Figure 6). |
| | Random Forest | The random forest splits the nodes of preprocessed data and then selects the split which results in homogeneous subnodes. The creation of subnodes like temperature, atmospheric pressure, and location magnitude shows increases in the homogeneity of resultant subnodes like longitude, latitude, and depth showing that leaves of the tree are falling, there is more water movement, more water bodies are high recorded on the earthquake prone locations, the range of magnitude is from 4.1 to 5.14, animal behaviour is more active, temperature is cold, time is early morning or late night, and variation in depth is negligible which shows high possibility of occurrences of earthquake (refer to Figure 7). |
|
|
