This paper comparing a method to automatically build a sentiment lexicon, with four well-known sentiment lexicons. For this purpose, an indirect evaluation is carried out. The lexicons are integrated into supervised sentiment classifiers and their performance is evaluated in two sentiment classification tasks in order to identify i) the most negative vs. not most negative opinions, and ii) the most positive vs. not most positive. Moreover, a set of textual features is integrated into the classifiers so as to analyze how these textual features improve the lexicon performance.