Force plots are a very intuitive and visual way to understand the influence of each feature on the models prediction for a specific instance of the data.
The Force plot, much like Feature importance, is based on a game theory method - Shapley values. Unlike many other methods, the Shapley method has a strong theoretical basis. In short, the features are treated as players that can form a coalition and play games. The outcome of the game is the prediction of the model. The importance of each feature is it’s average contribution to the different coalitions in comparison to the average prediction across all instances. The exact calculation of feature importance based on Shapley value is computationally inefficient, hence we use an approximation to this method.
The Force plot shows the influence of each feature on the current prediction. The base value is the averaged predicted probability across all samples. The green arrows represent the features that drive the prediction for this specific sample up, or in other words the values of this features have a positive influence on the prediction. The red arrows represent the features that drive the prediction for this specific sample down, or in other words the values of this features have a negative influence on the prediction. The bold value is the actual prediction for this sample. Note, that the usage of “positive” and “negative” is not in the context of “better/worse” but in the context of direction or proximity towards 1.
To learn more about shapely values, please refer to https://christophm.github.io/interpretable-ml-book/shapley.html