POS-RS: A Random Subspace method for sentiment classification based on part-of-speech analysis

Wang, Gang; Zhang, Zhu; Sun, Jianshan; Yang, Shanlin; Larson, Catherine

doi:10.1016/j.ipm.2014.09.004

Cited by 73 publications

(23 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We used terms (unigrams) as input features in our experiments. However, other features like n-grams (Vinodhini and Chandrasekaran, 2017), Part-of-Speech (Wang et al, 2015), Joint Sentiment Topic (He et al, 2011) or improvements in the quality of features (Xia et al, 2016) could also open new possibilities of investigation.…”

Section: Discussionmentioning

confidence: 99%

Unbalanced sentiment classification: an assessment of ANN in the context of sampling the majority class

Moraes

Valiati²,

Neto³

2018

Preprint

View full text Add to dashboard Cite

Many people make their opinions available on the Internet nowadays, and researchers have been proposing methods to automate the task of classifying textual reviews as positive or negative. Usual supervised learning techniques have been adopted to accomplish such a task. In practice, positive reviews are abundant in comparison to negative's. This context poses challenges to learning-based methods and data undersampling/oversampling are popular preprocessing techniques to overcome the problem. A combination of sampling techniques and learning methods, like Artificial Neural Networks (ANN) or Support Vector Machines (SVM), has been successfully adopted as a classification approach in many areas, while the sentiment classification literature has not explored ANN in studies that involve sampling methods to balance data. Even the performance of SVM, which is widely used as a sentiment learner, has been rarely addressed under the context of a preceding sampling method. This paper addresses document-level sentiment analysis with unbalanced data and focus on empirically assessing the performance of ANN in the context of undersampling the (majority) set of positive reviews. We adopted the performance of SVM as a baseline, since some studies have indicated SVM as being less subject to the class imbalance problem. Results are produced in terms of a traditional bag-of-words model with popular feature selection and weighting methods. Our experiments indicated that SVM are more stable than ANN in highly unbalanced (80%) data scenarios. However, under the discarding of information generated by random undersampling, ANN outperform SVM or produce comparable results.

show abstract

Section: Discussionmentioning

confidence: 99%

Unbalanced sentiment classification: an assessment of ANN in the context of sampling the majority class

Moraes

Valiati²,

Neto³

2018

Preprint

View full text Add to dashboard Cite

show abstract

“…After that, the majority voting is utilized as an ensemble strategy. Random Subspace [3,16,18,19,25,27,13,33,37] exploits fairly simple randomness approach for the feature selection. Training is done with a subset of the original feature space instead of including all features for each base learner in the ensemble.…”

Section: Ensemble Of Classifiersmentioning

confidence: 99%

Extended Feature Spaces Based Classifier Ensembles for Sentiment Analysis of Short Texts

Kilimci

Omurca

2018

ITC

View full text Add to dashboard Cite

Sentiment classification has become very popular to analyze opinions about events, products, and so on, especially for social networks such as Twitter. Due to the size limitation of expressing ideas on social networks, the classification performance needs to be boosted by proposing various techniques. In this work, the enhancement of feature space with word embedding based features is proposed to deal with the size limitation issues and the classification success of sentiment analysis is improved by employing classifier ensembles. The contributions of this paper are fivefold. First, the representative capabilities of features are enriched by using a semantic word embedding model and followingly the conventional feature selection techniques are compared. Second, traditional machine learning algorithms, namely naïve Bayes, support vector machine, and random forest are carried out to select baseline classifier for the proposed ensemble system. Third, three ensemble strategies namely, bagging, boosting, and random subspace are introduced to ensure the diversity of ensemble learning. Fourth, experiments are conducted to compare the performance of the models with the word embedding baseline. Eventually, a wide range of comparative experiments on Twitter datasets demonstrate that the classification performance of the proposed model significantly outperforms the state-of-the-art studies.

show abstract

“…Ensemble algorithms used in this paper are briefly mentioned. Bagging [1,12,[15][16][17][18][19]: Bagging produces new training dataset (bootstrap samples) from the original dataset by using replacement. In other words, the multiple versions are composed thereby performing bootstrap replicates of the training set and employing them as new training sets.…”

Section: Introductionmentioning

confidence: 99%

“…Then, the classifier is constituted on each of these samples and associated them with majority voting. Random Subspace [1,[16][17][18][20][21][22]: The idea behind of this approach is quite simple. Random subspace method aims to train with a subset of the original feature space instead of using extended version.…”

Section: Introductionmentioning

confidence: 99%

The Impact of Enhanced Space Forests with Classifier Ensembles on Biomedical Dataset Classification

Kilimci¹

2018

ijisae

View full text Add to dashboard Cite

In this paper, we propose to advance the classification success of classifier ensembles by investigating the contribution of enhanced space forests. For this purpose, this study especially is focused on enhanced feature spaces by implementing the most popular feature selection techniques, namely information gain, and chi-square. After performing these methods on the original feature space, training phase is evaluated with all the original and the modified versions of most significant features, which are acquired by applying difference operator to the original features and the selected features with feature selection methods. That is, the new training dataset is constructed by combining the original features and the new ones. Then, the training is done with the well-known classification algorithm namely, decision tree using the enhanced feature space. Finally, three types of ensemble algorithms namely, bagging, random subspace, and random forest are carried out. A wide range of comparative experiments are conducted on publicly available and widely-used 36 datasets from the UCI machine learning repository to observe the impact of the enhanced space forests with classifier ensembles. Experiment results demonstrate that the proposed enhanced space forests perform better classification accuracy than the state of the art studies. Approximately, 1%-3% improvement of the classification success is an indicator that our proposed technique is efficient.

show abstract

POS-RS: A Random Subspace method for sentiment classification based on part-of-speech analysis

Cited by 73 publications

References 52 publications

Unbalanced sentiment classification: an assessment of ANN in the context of sampling the majority class

Unbalanced sentiment classification: an assessment of ANN in the context of sampling the majority class

Extended Feature Spaces Based Classifier Ensembles for Sentiment Analysis of Short Texts

The Impact of Enhanced Space Forests with Classifier Ensembles on Biomedical Dataset Classification

Contact Info

Product

Resources

About