Highlight words which explains a prediction.

plot_text_explanations(explanations, ...)

Arguments

explanations

object returned by the lime.character function.

...

parameters passed to sizingPolicy

See also

Other explanation plots: plot_explanations, plot_features

Examples

# We load a precalculated explanation set based on the procedure in the ?lime # examples explanations <- .load_text_example() # We see that the explanations are in the expected format print(explanations)
#> model_type case label label_prob model_r2 model_intercept feature #> in classification 1 1 0.6418385 0.9879563 0.3321225 in #> we classification 1 1 0.6418385 0.9879563 0.3321225 we #> in1 classification 2 1 0.8022363 0.8551634 0.3210651 in #> We classification 2 1 0.8022363 0.8551634 0.3210651 We #> those classification 3 0 0.5432571 0.8868849 0.6909923 those #> we1 classification 3 0 0.5432571 0.8868849 0.6909923 we #> in2 classification 4 1 0.8719526 0.6569822 0.4941383 in #> Section classification 4 1 0.8719526 0.6569822 0.4941383 Section #> of classification 5 1 0.7316587 0.3242455 0.3535088 of #> our classification 5 1 0.7316587 0.3242455 0.3535088 our #> feature_value feature_weight feature_desc #> in in -0.03694821 in #> we we 0.34825310 we #> in1 in -0.01236806 in #> We We 0.42784027 We #> those those 0.07281582 those #> we1 we -0.26484904 we #> in2 in 0.01006400 in #> Section Section 0.33433000 Section #> of of 0.01377475 of #> our our 0.28981386 our #> data #> in we validate them in section and conclude in section #> we we validate them in section and conclude in section #> in1 We developed a two-level machine learning approach that in the first level considers two different properties important for protein-protein binding derived from structural models of V3 and V3 sequences. #> We We developed a two-level machine learning approach that in the first level considers two different properties important for protein-protein binding derived from structural models of V3 and V3 sequences. #> those specifically we hypothesize that the judgments of those individuals high in this trait will be more influenced by previously presented anchors whereas those individuals low in this trait will be less influenced by the anchor #> we1 specifically we hypothesize that the judgments of those individuals high in this trait will be more influenced by previously presented anchors whereas those individuals low in this trait will be less influenced by the anchor #> in2 Finally, in Section we discuss the results and give conclusions #> Section Finally, in Section we discuss the results and give conclusions #> of We also illustrate the application of our SYMBOL -mixing generalization bounds to general classes of learning algorithms, including Support Vector Regression (SVR) CITATION , Kernel Ridge Regression CITATION , and Support Vector Machines (SVMs) CITATION #> our We also illustrate the application of our SYMBOL -mixing generalization bounds to general classes of learning algorithms, including Support Vector Regression (SVR) CITATION , Kernel Ridge Regression CITATION , and Support Vector Machines (SVMs) CITATION #> prediction #> in 0.6418385, 0.3581615 #> we 0.6418385, 0.3581615 #> in1 0.8022363, 0.1977637 #> We 0.8022363, 0.1977637 #> those 0.4567429, 0.5432571 #> we1 0.4567429, 0.5432571 #> in2 0.8719526, 0.1280474 #> Section 0.8719526, 0.1280474 #> of 0.7316587, 0.2683413 #> our 0.7316587, 0.2683413
# We can now get the explanations in the context of the input text plot_text_explanations(explanations)