Explain me about this Support Vector Machine training set graph?

Chloronic

I made a support vector machine classification with supplied training set of VADER labelling for positive and negative sentiment (1000 positive sentiment and 400 negative sentiment). I don't understand why the graph for positive sentiment is negative on the function value? and why my negative sentiment is having a negative function value on the positive sentiment? Will it affect the accuracy on evaluation model? Although I already set the positive class manually in the Performance (Binomial Classification) Operator. Here is the Kernel Dot Model graph:


And here is my classification process:


P.S.: VADER labelling process is already done on the Read Excel Operator and Split Data with 70% or 1400 data training set + 30% or 600 test set.