LIME - Understanding its Output for Regression

[SorenSc]

Normally, I'm writing questions on Stackoverflow; unfortunately, often nobody replies to these more or less specific questions. If my question does not belong here, please be gentle and point me where I should ask it. I'll act accordingly :)

I couldn't find the documentation that describes explanation.local_pred or explanation.predicted_value. What is the meaning of the two in case of a regression?

I was expecting that explanation.intercept + x_1 * explanation.local_exp[1][1][1] + x_2 * explanation.local_exp[1][0][1] results in explanation.local_pred but that doesn't seem to be the case. Here, I used a regression problem with two input features---x_1 and x_2---for illustration.

Btw: Why do I have to explanation.local_exp[1][**1**][1] when referring to the explanation for x_1 and explanation.local_exp[1][**0**][1] for x_2? Why is the order of the variables in the list of local explanation reversed? Isn't that a little unintuitive?

Thank you very much for your help!