No abstract

Twitter sentiment analysis is a challenging problem in natural language processing. For this purpose, supervised learning techniques have mostly been employed, which require labeled data for training. However, it is very time consuming to label datasets of large size. To address this issue, unsupervised learning techniques such as clustering can be used. In this study, we explore the possibility of using hierarchical clustering for twitter sentiment analysis. Three hierarchical-clustering techniques, namely single linkage (SL), complete linkage (CL) and average linkage (AL), are examined. A cooperative framework of SL, CL and AL is built to select the optimal cluster for tweets wherein the notion of optimal-cluster selection is operationalized using majority voting. The hierarchical clustering techniques are also compared with k-means and two state-of-the-art classifiers (SVM and Naïve Bayes). The performance of clustering and classification is measured in terms of accuracy and time efficiency. The experimental results indicate that cooperative clustering based on majority voting approach is robust in terms of good quality clusters with tradeoff of poor time efficiency. The results also suggest that the accuracy of the proposed clustering framework is comparable to classifiers which is encouraging. INDEX TERMS Cooperative clustering, majority voting, sentiment analysis, twitter sentiment analysis.

Twitter sentiment analysis is a challenging task that involves various preprocessing steps including dimensionality reduction. Dimensionality reduction helps ensure low computational complexity and performance improvement during the classification process. In Twitter data, each tweet has feature values which may or may not reflect a person's response. Therefore, a large number of sparse data points are generated when tweets are represented as feature matrix, eventually increasing computational overheads and error rates in Twitter sentiment analysis. This study proposes a novel preprocessing technique called class association and attribute relevancy based imputation algorithm (CAARIA) to reduce the Twitter data size. CAARIA achieves the dimensionality reduction goal by imputing those tweets that belong to the same class and also share useful information. The performance of two classifiers (Naïve Bayes and support vector machines) is evaluated on three Twitter datasets in terms of classification accuracy, measured as area under curve, and time efficiency. CAARIA is also compared against two widely used feature selection (dimensionality reduction) techniques, information gain (IG) and Pearson's correlation (PC). The findings reveal that CAARIA outperforms IG and PC in terms of classification accuracy and time efficiency. These results suggest that CAARIA is a robust data preprocessing technique for the classification task.

Maintenance of architectural documentation is a prime requirement for evolving software systems. New versions of software systems are launched after making the changes that take place in a software system over time. The orphan adoption problem, which deals with the issue of accommodation of newly introduced resources (orphan resources) in appropriate subsystems in successive versions of a software system, is a significant problem. The orphan adoption algorithm has been developed to address this problem. For evolving software systems, it would be useful to recover the architecture of subsequent versions of a software system by using existing architectural information. In this paper, we explore supervised learning techniques (classifiers) for recovering the architecture of subsequent versions of a software system by taking benefit of existing architectural information. We use three classifiers, i.e., Bayesian classifier, k-Nearest Neighbor classifier and Neural Network for orphan adoption. We conduct experiments to compare the performance of the classifiers using various dependencies between entities in a software system. Our experiments highlight correspondence between the orphan adoption algorithm and the classifiers, and also reveal their strengths and weaknesses. To combine strengths of individual classifiers, we propose using a multiclassifier approach in which classifiers work cooperatively to improve classification accuracy. Experiments show that there is significant improvement in results when our proposed multiclassifier approach is used.