CS 229 Final Project Sentiment Causation Extraction
نویسندگان
چکیده
In state of the art sentiment analysis, text is analyzed for a single unidimensional sentiment or opinion score. This unidimensional sentiment, however, cannot capture the nuances of emotions: for example, two texts that respectively convey anger and sadness will both have a negative sentiment associated with them, while carrying very different connotations. Thus, we require a tool that allows more sophisticated analysis of emotional text. This need for more sophisticated relation extraction is relevant outside of sentiment analysis as well. For example, causal relation extraction has important applications for building question and answer systems. Relation extraction remains an incredibly difficult problem to solve, but offers many promises such as constructing abstract knowledge representations or constructing narrative sequences from newspapers [1]. In this work, we propose building a causal relation extraction tool specifically for extracting the cause of emotions. Although we focus on causal extraction of emotions, the tool is generalizable to causal extraction in other domains.
منابع مشابه
CS 229 Project: Using Vector Representations to Augment Sentiment Analysis Training Data
While the accuracy of supervised sentiment classification algorithms has steadily increased in recent years, acquiring new human-labeled sentiment data remains expensive. In this paper, we explore the effectiveness of increasing the training data set size for sentiment classification algorithms by adding unlabeled phrases whose sentiments are inferred by their proximity to labeled training phra...
متن کاملCS 224D Final Project: Neural Network Ensembles for Sentiment Classification
We investigate the effect of ensembling on two simple models: LSTM and bidirectional LSTM. These models are used for fine-grained sentiment classification on the Stanford Sentiment Treebank dataset. We observe that ensembling improves the classification accuracy by about 3% over single models. Moreover, the more complex model, bidirectional LSTM, benefits more from ensembling.
متن کاملCS 229 = = Final Project Report SPEECH & NOISE SEPARATION
In this course project I investigated machine learning approaches on separating speech signals from background noise. Keywords—MFCC, SVM, noise separation, source separation, spectrogram
متن کاملCS 224N Project Building an Emotional Relation Extraction Tool
Unidimensional sentiment or opinion analysis cannot capture the differences between similarly valenced (positive/negative) emotions (such as angry and sad), and we argue the need for more sophisticated emotion analysis tools that go beyond classification into valenced or basic emotion categories. In particular, we propose that building a relation extraction tool that can extract the cause of a ...
متن کامل