Evaluation of the Maximum Entropy Method

In this subsection we present the results of our evaluation over the training and testing data of the SENSEVAL-2 Spanish lexical-sample task. This corpus was parsed using Conexor Functional Dependency Grammar parser for Spanish Tapanainen97.

The classifiers were built from the training data and evaluated over the test data. Table 7 shows which combination of groups of features works better for every POS and which work better for all words together.

Table: Baseline: Accuracy Results of Applying ME on SENSEVAL-2 Spanish Data

Accuracy Feature selection


0.683 LWSBCk5
Verbs 0.595 sk5
Adjectives 0.783 LWsBCp
ALL 0.671 0LWSBCk5

This work entails an exhaustive search looking for the most accurate combination of features. The values presented here are merely informative and indicate the maximum accuracy that the system can achieve with a particular set of features.