Weighted Vote-Based Classifier Ensemble for Named Entity Recognition

Abstract

In this article, we report the search capability of Genetic Algorithm (GA) to construct a weighted vote-based classifier ensemble for Named Entity Recognition (NER). Our underlying assumption is that the reliability of predictions of each classifier differs among the various named entity (NE) classes. Thus, it is necessary to quantify the amount of voting of a particular classifier for a particular output class. Here, an attempt is made to determine the appropriate weights of voting for each class in each classifier using GA. The proposed technique is evaluated for four leading Indian languages, namely Bengali, Hindi, Telugu, and Oriya, which are all resource-poor in nature. Evaluation results yield the recall, precision and F-measure values of 92.08%, 92.22%, and 92.15%, respectively for Bengali; 96.07%, 88.63%, and 92.20%, respectively for Hindi; 78.82%, 91.26%, and 84.59%, respectively for Telugu; and 88.56%, 89.98%, and 89.26%, respectively for Oriya. Finally, we evaluate our proposed approach with the benchmark dataset of CoNLL-2003 shared task that yields the overall recall, precision, and F-measure values of 88.72%, 88.64%, and 88.68%, respectively. Results also show that the vote based classifier ensemble identified by the GA-based approach outperforms all the individual classifiers, three conventional baseline ensembles, and some other existing ensemble techniques. In a part of the article, we formulate the problem of feature selection in any classifier under the single objective optimization framework and show that our proposed classifier ensemble attains superior performance to it.

Joint model for feature selection and parameter optimization coupled with classifier ensemble in chemical mention recognition

“…The solutions that exist on a particular population do not conform to the uniform characteristics. In order to further improve the performance we combine all the CRF and SVM based models using a SOO based classifier ensemble technique [6]. Overall evaluation results along with the baseline models are reported in Table 3 After application of the MOO based feature selection and parameter optimization technique for the CRF based classifier we obtain a set of Pareto optimal solutions.…”

“…Some of the solutions on the best population may have high recall values whereas some could have high precision values. Thus instead of selecting a single solution we use a SOO based classifier ensemble technique [6] to combine the solutions, obtained in the best population.…”

“…The SOO and MOO based techniques are based on GA and NSGA-II, respectively. For SOO and MOO based ensemble learning we employ the algorithms proposed in [6,24], respectively. We use real encoding, i.e.…”

“…6 Brief statistics of the datasets are reported in Table 2. The training set was collected from the Medline abstracts.…”

“…The feature selection is performed for both the classifiers, CRF and SVM. The set of solutions obtained in the final population of these two after application of GA based technique are then merged together using a GA based classifier ensemble technique [6]. Similarly after the application of MOO based technique we obtain a set of solutions on the final Pareto optimal front for CRF and SVM each.…”