Quality of Word Embeddings on Sentiment Analysis Tasks

2018

TASS 2018 workshop proposes different challenges on semantic analysis in Spanish. This work presents our participation as team Atalaya in the task of polarity classification of tweets. We followed standard techniques in preprocessing, representation and classification, and also explored some novel ideas. In particular, to obtain tweet embeddings we trained subword-aware word embeddings and use a weighted scheme to average them. To deal with overfitting problems caused by training data scarcity, we tried a data augmentation strategy based on two-way machine translation. Experiments with linear classifiers and neural models show competitive results for the different subtasks proposed in the challenge.

2016

A key point in Sentiment Analysis is to determine the polarity of the sentiment implied by a certain word or expression. In basic Sentiment Analysis systems this sentiment polarity of the words is accounted and weighted in different ways to provide a degree of positivity/negativity. Currently words are also modelled as continuous dense vectors, known as word embeddings, which seem to encode interesting semantic knowledge. With regard to Sentiment Analysis, word embeddings are used as features to more complex supervised classification systems to obtain sentiment classifiers. In this paper we compare a set of existing sentiment lexicons and sentiment lexicon generation techniques. We also show a simple but effective technique to calculate a word polarity value for each word in a domain using existing continuous word embeddings generation methods. Further, we also show that word embeddings calculated on in-domain corpus capture the polarity better than the ones calculated on general-do...

The International AAAI Conference on Web and Social Media, 2017

A word embedding is a low-dimensional, dense and real-valued vector representation of a word. Word embeddings have been used in many NLP tasks. They are usually generated from a large text corpus. The embedding of a word captures both its syntactic and semantic aspects. Tweets are short, noisy and have unique lexical and semantic features that are different from other types of text. Therefore, it is necessary to have word embeddings learned specifically from tweets. In this paper, we present ten word embedding data sets. In addition to the data sets learned from just tweet data, we also built embedding sets from the general data and the combination of tweets with the general data. The general data consist of news articles, Wikipedia data and other web data. These ten embedding models were learned from about 400 million tweets and 7 billion words from the general text. In this paper, we also present two experiments demonstrating how to use the data sets in some NLP tasks, such as tweet sentiment analysis and tweet topic classification tasks.

International Journal of Web-Based Learning and Teaching Technologies

Sentiment analysis is the practice of eliciting a sentiment orientation of people's opinions (i.e. positive, negative and neutral) toward the specific entity. Word embedding technique like Word2vec is an effective approach to encode text data into real-valued semantic feature vectors. However, it fails to preserve sentiment information that results in performance deterioration for sentiment analysis. Additionally, big sized textual data consisting of large vocabulary and its associated feature vectors demands huge memory and computing power. To overcome these challenges, this research proposed a MapReduce based Sentiment weighted Word2Vec (MSW2V), which learns the sentiment and semantic feature vectors using sentiment dictionary and big textual data in a distributed MapReduce environment, where memory and computing power of multiple computing nodes are integrated to accomplish the huge resource demand. Experimental results demonstrate the outperforming performance of the MSW2V c...

WorldCist'18 - 6th World Conference on Information Systems and Technologies, Napoli, Italy, 2018

Dense and low dimensional word embeddings opened up the possibility to analyze text polarity with highly successful deep learning techniques like Convolution Neural Networks. In this paper we utilize pretrained word vectors in combination with simple neural networks of stacked convolution and max-pooling layers, to explore the role of dataset size and document length in sentiment polarity prediction. We experiment with song lyrics and reviews of products or movies and see that convolution-pooling combination is very fast and yet quiet effective. We also find interesting relations between dataset size, text length and length of feature maps with classification accuracy. Our next goal is the design of a generic neural architecture for analyzing polarity of various text types, with high accuracy and few hyper-parameter changes.

viXra, 2020

Sentiment Analysis is one of the primary areas of natural language processing and information retrieval being tackled by researchers to date, and for good reason; the internet. The internet is a mostly untapped source of rich amounts of data that can be used to gauge the opinions of people, in reference to any number of topics. Twitter is one such platform designed for people to voice their opinions in the form of tweets about any topic they desire. My project will set out to investigate the best way to be able to analyse the sentiment of these aforementioned tweets using machine learning techniques. I will be training word vector-based, and paragraph vector-based models on a dataset consisting of 1.6 million tweets, in conjunction with various classifiers in order to find the best performing method in which to obtain the sentiment of tweets.

Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), 2016

This paper presents our approach for SemEval 2016 task 4: Sentiment Analysis in Twitter. We participated in Subtask A: Message Polarity Classification. The aim is to classify Twitter messages into positive, neutral, and negative polarity. We used a lexical resource for pre-processing of social media data and train a neural network model for feature representation. Our resource includes dictionaries of slang words, contractions, abbreviations, and emoticons commonly used in social media. For the classification process, we pass the features obtained in an unsupervised manner into an SVM classifier.

Bulletin of Electrical Engineering and Informatics, 2025

In this study, we apply various deep learning methods incorporating word embedding features to evaluate their impact on improving classification performance in sentiment analysis. The methods employed include conditional random field (CRF), bidirectional long short term memory (BLSTM), and convolutional neural network (CNN). Our experiments utilize social media data from restaurant review. By testing different iterations of these deep learning techniques with various word embedding features, we found that the BLSTM algorithm achieved the highest accuracy of 80.00% before integrating word embedding features. After incorporating word embeddings, the BLSTM with the word2vec feature achieved an accuracy of 87.00%. Notably, the CNN showed a significant improvement with the FastText feature. Considering all evaluation metrics-accuracy, precision, recall, and F1-score-the BLSTM algorithm consistently demonstrated the best performance across different word embeddings.

Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval-2017)

Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing, 2019

Social media are providing the humus for the sharing of knowledge and experiences and the growth of community activities (e.g., debating about different topics). The analysis of the user-generated content in this area usually relies on Sentiment Analysis. Word embeddings and Deep Learning have attracted extensive attention in various sentiment detection tasks. In parallel, the literature exposed the drawbacks of traditional approaches when content belonging to specific contexts is processed with general techniques. Thus, ad-hoc solutions are needed to improve the effectiveness of such systems. In this paper, we focus on user-generated content coming from the e-learning context to demonstrate how distributional semantic approaches trained on smaller context-specific textual resources are more effective with respect to approaches trained on bigger general-purpose ones. To this end, we build context-trained embeddings from online course reviews using state-of-the-art generators. Then, those embeddings are integrated in a deep neural network we designed to solve a polarity detection task on reviews in the e-learning context, modeled as a regression. By applying our approach on embeddings trained using background corpora from different contexts, we show that the performance is better when the background context is aligned with the regression context.