SSN MLRG1 at SemEval-2018 Task 1: Emotion and Sentiment Intensity Detection Using Rule Based Feature Selection

Proceedings of the 12th International Workshop on Semantic Evaluation (SemEval-2018), pages 324–328 New Orleans, Louisiana, June 5–6, 2018. ©2018 Association for Computational Linguistics SSN MLRG1 at SemEval-2018 Task 1: Emotion and Sentiment Intensity Detection Using Rule Based Feature Selection Angel Deborah S, Rajalakshmi S, S Milton Rajendram, Mirnalinee T T SSN College of Engineering Chennai 603 110, Tamil Nadu, India angeldeborahs@ssn.edu.in, rajalakshmis@ssn.edu.in, miltonrs@ssn.edu.in, mirnalineett@ssn.edu.in Abstract The system developed by the SSN MLRG1 team for Semeval-2018 task 1 on affect in tweets uses rule based feature selection and one-hot encoding to generate the input fea- ture vector. Multilayer Perceptron was used to build the model for emotion intensity ordinal classification, sentiment analysis ordinal clas- sification and emotion classfication subtasks. Support Vector Regression was used to build the model for emotion intensity regression and sentiment intensity regression subtasks. 1 Introduction Twitter is a huge microblogging service with more than 500 million tweets per day from differ- ent locations of the world and in different lan- guages (Saif and Felipe, 2017). Tweets are of- ten used to convey ones emotions, opinions to- wards products, and stance over issues (Nabil et al., 2016). Automatically detecting emotion intensities in tweets has several applications, in- cluding commerce (Jansen et al., 2009), crisis management (Verma et al., 2011), tracking brand and product perception, tracking support for is- sues and policies, and tracking public health and well-being (Chew and Eysenbach, 2010). The task is challenging because of the informal writing style, the semantic diversity of content as well as the “unconventional” grammar. These challenges in building a classification model and regression model can be handled by using proper approaches to feature generation and machine learning. There are several machine learning techniques that can be used for sentiment intensity predic- tion or emotion intensity prediction. Some of the approaches inlclude Artificial Neural Network (ANN) (Sudipta et al., 2017), Random Forests, Support Vector Machine (SVM), Naive Bayes (NB) (Tabari et al., 2017), Multi-Kernel Gaussian Process (MKGP) (Angel Deborah et al., 2017a,b), AdaBoost Regressor (ABR), Bagging Regres- sor (BR)(Jiang et al., 2017) and Deep Learning (DL)(Pivovarova et al., 2017). 2 Multi-Layer Perceptron A Multilayer Perceptron (MLP) as shown in Fig- ure 1 is a feed-forward Neural Network model that maps input data sets onto appropriate output sets. An MLP has many layers of nodes in a directed graph, with each layer connected to the next layer. A neuron is a processing element with activation Input1 Input2 Input3 Output1 Output2 Figure 1: Multi-Layer Perceptron. function (in the input layer the activation func- tion is not applied). The output layer has as many nuerons as the number of class labels in the prob- lem. Each connection has a weight assigned to it. Output of each neuron is calculated by apply- ing the activation function on the weighted sum of the inputs. Linear, sigmoid, tanh, elu, soft- plus, softmax and relu are some of the commonly used activation functions. The supervised learning problem of the MLP can be solved with the back- propagation algorithm …

Content truncated to 500 words