Word embeddings that provide continuous low-dimensional vector representations of words have been extensively used for various natural language processing tasks. However, existing context-based word embeddings such as Word2vec and GloVe typically fail to capture sufficient sentiment information, which may result in words with similar vector representations having an opposite sentiment polarity (e.g., good and bad), thus degrading sentiment analysis performance. To tackle this problem, recent studies have suggested learning sentiment embeddings to incorporate the sentiment polarity (positive and negative) information from labeled corpora. This study adopts another strategy to learn sentiment embeddings. Instead of creating a new word embedding from labeled corpora, we propose a word vector refinement model to refine existing pretrained word vectors using real-valued sentiment intensity scores provided by sentiment lexicons. The idea of the refinement model is to improve each word vector such that it can be closer in the lexicon to both semantically and sentimentally similar words (i.e., those with similar intensity scores) and further away from sentimentally dissimilar words (i.e., those with dissimilar intensity scores). An obvious advantage of the proposed method is that it can be applied to any pretrained word embeddings. In addition, the intensity scores can provide more fine-grained (real-valued) sentiment information than binary polarity labels to guide the refinement process. Experimental results show that the proposed refinement model can improve both conventional word embeddings and previously proposed sentiment embeddings for binary, ternary, and fine-grained sentiment classification on the SemEval and Stanford Sentiment Treebank datasets.

Dimensional sentiment analysis aims to recognize continuous numerical values in multiple dimensions such as the valence-arousal (VA) space. Compared to the categorical approach that focuses on sentiment classification such as binary classification (i.e., positive and negative), the dimensional approach can provide a more fine-grained sentiment analysis. This study proposes a tree-structured regional CNN-LSTM model consisting of two parts: regional CNN and LSTM to predict the VA ratings of texts. Unlike a conventional CNN which considers a whole text as input, the proposed regional CNN uses a part of the text as a region, dividing an input text into several regions such that the useful affective information in each region can be extracted and weighted according to their contribution to the VA prediction. Such regional information is sequentially integrated across regions using LSTM for VA prediction. By combining the regional CNN and LSTM, both local (regional) information within sentences and long-distance dependencies across sentences can be considered in the prediction process. To further improve performance, a region division strategy is proposed to discover task-relevant phrases and clauses to incorporate structured information into VA prediction. Experimental results on different corpora show that the proposed method outperforms lexicon-, regression-, conventional NN and other structured NN methods proposed in previous studies.

Massive open online courses (MOOCs) have recently gained worldwide attention from educational institutes. MOOCs provide a new option for learning, yet measurable learning benefits of MOOCs still need to be investigated. Collecting data of three MOOCs at Yuan Ze University (YZU), this paper intended to classify learning behaviors among 1489 students on the MOOC platform at YZU. This study further examined learning outcomes in MOOCs by different types of learners. The Ward’s hierarchical and k-means non-hierarchical clustering methods were employed to classify types of learners’ behavior while they engaged in learning activities on the MOOC platform. Three types of MOOC learners were classified—active learner, passive learner, and bystander. Active learners who submitted assignments on time and frequently watched lecture videos showed a higher completion rate and a better grade in the course. MOOC learners who participated in online discussion forum reported a higher rate of passing the course and a better score than those inactive classmates. The finding of this study suggested that the first 2 weeks was a critical point of time to retain students in MOOCs. MOOC instructors need to carefully design course and detect risk behaviors of students in early of the classes to prevent students from dropping out of the course. The feature design of discussion forum is to provide peer interaction and facilitate online learning. Our results suggested that timely feedback by instructors or facilitators on discussion forum could enhance students’ engagement in MOOCs.

Machine learning is useful for grid-based crime prediction. Many previous studies have examined factors including time, space, and type of crime, but the geographic characteristics of the grid are rarely discussed, leaving prediction models unable to predict crime displacement. This study incorporates the concept of a criminal environment in grid-based crime prediction modeling, and establishes a range of spatial-temporal features based on 84 types of geographic information by applying the Google Places API to theft data for Taoyuan City, Taiwan. The best model was found to be Deep Neural Networks, which outperforms the popular Random Decision Forest, Support Vector Machine, and K-Near Neighbor algorithms. After tuning, compared to our design’s baseline 11-month moving average, the F1 score improves about 7% on 100-by-100 grids. Experiments demonstrate the importance of the geographic feature design for improving performance and explanatory ability. In addition, testing for crime displacement also shows that our model design outperforms the baseline.

An increasing amount of research has recently focused on dimensional sentiment analysis that represents affective states as continuous numerical values on multiple dimensions, such as valence-arousal (VA) space. Compared to the categorical approach that represents affective states as distinct classes (e.g., positive and negative), the dimensional approach can provide more fine-grained (real-valued) sentiment analysis. However, dimensional sentiment resources with valence-arousal ratings are very rare, especially for the Chinese language. Therefore, this study aims to: (1) Build a Chinese valence-arousal resource called Chinese EmoBank, the first Chinese dimensional sentiment resource featuring various levels of text granularity including 5,512 single words, 2,998 multi-word phrases, 2,582 single sentences, and 2,969 multi-sentence texts. The valence-arousal ratings are annotated by crowdsourcing based on the Self-Assessment Manikin (SAM) rating scale. A corpus cleanup procedure is then performed to improve annotation quality by removing outlier ratings and improper texts. (2) Evaluate the proposed resource using different categories of classifiers such as lexicon-based, regression-based, and neural-network-based methods, and comparing their performance to a similar evaluation of an English dimensional sentiment resource.