Search results for: patent sentiment analysis

Authors: Renukswamy Chikkamath, Vishvapalsinhji Ramsinh Parmar, Christoph Hewel, Markus Endres

Abstract:

Given a patent document, identifying distinct semantic annotations is an interesting research aspect. Text annotation helps the patent practitioners such as examiners and patent attorneys to quickly identify the key arguments of any invention, successively providing a timely marking of a patent text. In the process of manual patent analysis, to attain better readability, recognising the semantic information by marking paragraphs is in practice. This semantic annotation process is laborious and time-consuming. To alleviate such a problem, we proposed a dataset to train machine learning algorithms to automate the highlighting process. The contributions of this work are: i) we developed a multi-class dataset of size 150k samples by traversing USPTO patents over a decade, ii) articulated statistics and distributions of data using imperative exploratory data analysis, iii) baseline Machine Learning models are developed to utilize the dataset to address patent paragraph highlighting task, and iv) future path to extend this work using Deep Learning and domain-specific pre-trained language models to develop a tool to highlight is provided. This work assists patent practitioners in highlighting semantic information automatically and aids in creating a sustainable and efficient patent analysis using the aptitude of machine learning.

Keywords: machine learning, patents, patent sentiment analysis, patent information retrieval

Procedia PDF Downloads 61

Authors: Zane Turner, Kevin Labille, Susan Gauch

Abstract:

Sentiment analysis is a broad and expanding field that aims to extract and classify opinions from textual data. Lexicon-based approaches are based on the use of a sentiment lexicon, i.e., a list of words each mapped to a sentiment score, to rate the sentiment of a text chunk. Our work focuses on predicting stock price change using a sentiment lexicon built from financial conference call logs. We present a method to generate a sentiment lexicon based upon an existing probabilistic approach. By using a domain-specific lexicon, we outperform traditional techniques and demonstrate that domain-specific sentiment lexicons provide higher accuracy than generic sentiment lexicons when predicting stock price change.

Keywords: computational finance, sentiment analysis, sentiment lexicon, stock movement prediction

Procedia PDF Downloads 102

Authors: Zane Turner, Kevin Labille, Susan Gauch

Abstract:

Sentiment analysis is a broad and expanding field that aims to extract and classify opinions from textual data. Lexicon-based approaches are based on the use of a sentiment lexicon, i.e., a list of words each mapped to a sentiment score, to rate the sentiment of a text chunk. Our work focuses on predicting stock price change using a sentiment lexicon built from financial conference call logs. We introduce a method to generate a sentiment lexicon based upon an existing probabilistic approach. By using a domain-specific lexicon, we outperform traditional techniques and demonstrate that domain-specific sentiment lexicons provide higher accuracy than generic sentiment lexicons when predicting stock price change.

Keywords: computational finance, sentiment analysis, sentiment lexicon, stock movement prediction

Procedia PDF Downloads 144

Authors: Jie Liu, Xudong Luo, Pingping Lin, Yifan Fan

Abstract:

Facebook, Twitter, Weibo, and other social media and significant e-commerce sites generate a massive amount of online texts, which can be used to analyse people’s opinions or sentiments for better decision-making. So, sentiment analysis, especially fine-grained sentiment analysis, is a very active research topic. In this paper, we survey various methods for fine-grained sentiment analysis, including traditional sentiment lexicon-based methods, machine learning-based methods, and deep learning-based methods in aspect/target/attribute-based sentiment analysis tasks. Besides, we discuss their advantages and problems worthy of careful studies in the future.

Keywords: sentiment analysis, fine-grained, machine learning, deep learning

Procedia PDF Downloads 216

Authors: Pingping Lin, Xudong Luo

Abstract:

Natural language often conveys emotions of speakers. Therefore, sentiment analysis on what people say is prevalent in the field of natural language process and has great application value in many practical problems. Thus, to help people understand its application value, in this paper, we survey various applications of sentiment analysis, including the ones in online business and offline business as well as other types of its applications. In particular, we give some application examples in intelligent customer service systems in China. Besides, we compare the applications of sentiment analysis on Twitter, Weibo, Taobao and Facebook, and discuss some challenges. Finally, we point out the challenges faced in the applications of sentiment analysis and the work that is worth being studied in the future.

Keywords: application, natural language processing, online comments, sentiment analysis

Procedia PDF Downloads 227

Authors: Sannikumar Patel, Brian Nolan, Markus Hofmann, Philip Owende, Kunjan Patel

Abstract:

Sentiment analysis and opinion mining have become emerging topics of research in recent years but most of the work is focused on data in the English language. A comprehensive research and analysis are essential which considers multiple languages, machine translation techniques, and different classifiers. This paper presents, a comparative analysis of different approaches for multilingual sentiment analysis. These approaches are divided into two parts: one using classification of text without language translation and second using the translation of testing data to a target language, such as English, before classification. The presented research and results are useful for understanding whether machine translation should be used for multilingual sentiment analysis or building language specific sentiment classification systems is a better approach. The effects of language translation techniques, features, and accuracy of various classifiers for multilingual sentiment analysis is also discussed in this study.

Keywords: cross-language analysis, machine learning, machine translation, sentiment analysis

Procedia PDF Downloads 676

Authors: Chothmal, Basant Agarwal

Abstract:

Sentiment analysis means to classify a given review document into positive or negative polar document. Sentiment analysis research has been increased tremendously in recent times due to its large number of applications in the industry and academia. Sentiment analysis models can be used to determine the opinion of the user towards any entity or product. E-commerce companies can use sentiment analysis model to improve their products on the basis of users’ opinion. In this paper, we propose a new One-class Support Vector Machine (One-class SVM) based sentiment analysis model for movie review documents. In the proposed approach, we initially extract features from one class of documents, and further test the given documents with the one-class SVM model if a given new test document lies in the model or it is an outlier. Experimental results show the effectiveness of the proposed sentiment analysis model.

Keywords: feature selection methods, machine learning, NB, one-class SVM, sentiment analysis, support vector machine

Procedia PDF Downloads 478

Authors: Pingping Lin, Xudong Luo, Yifan Fan

Abstract:

Sentiment analysis is a very active research topic. Every day, Facebook, Twitter, Weibo, and other social media, as well as significant e-commerce websites, generate a massive amount of comments, which can be used to analyse peoples opinions or emotions. The existing methods for sentiment analysis are based mainly on sentiment dictionaries, machine learning, and deep learning. The first two kinds of methods rely on heavily sentiment dictionaries or large amounts of labelled data. The third one overcomes these two problems. So, in this paper, we focus on the third one. Specifically, we survey various sentiment analysis methods based on convolutional neural network, recurrent neural network, long short-term memory, deep neural network, deep belief network, and memory network. We compare their futures, advantages, and disadvantages. Also, we point out the main problems of these methods, which may be worthy of careful studies in the future. Finally, we also examine the application of deep learning in multimodal sentiment analysis and aspect-level sentiment analysis.

Keywords: document analysis, deep learning, multimodal sentiment analysis, natural language processing

Procedia PDF Downloads 129

Authors: Amaliny Yoganathan-Hasselbeck

Abstract:

Based on the assumption, that strategic licensing is more valuable and sustainable for the economy than a legal dispute and action for an injunction, the strategy of licensing in patent infringement cases was studied. A theoretical framework was developed based on the transaction costs approach, describing the major variables within the process of licensing to an alleged patent infringer. An exploratory case study analysis was conducted on the basis of expert interviews with patent licensing agencies, patent attorneys, licensing departments of companies and research institutions. Key findings define the major criteria in each step of the licensing process and include the factors determining the intensity of patent tracking e.g. patent policies, the decision criteria when dealing with patent infringement cases, e.g. market position and reputation, and the transaction itself starting with the initiation of the contact with the alleged patent infringer, negotiating the licensing contract and monitoring the license agreement.

Keywords: innovation, licensing, patent, patent infringement, strategy, technology

Procedia PDF Downloads 437

Authors: Qiang Bu

Abstract:

This paper compares the explanatory power of sentiment level and sentiment shock. The preliminary test results show that sentiment shock plays a more significant role in explaining stocks returns, including the raw return and abnormal return. We also find that sentiment shock beta has a higher statistical significance than sentiment beta. These finding sheds new light on the relationship between investor sentiment and stock returns.

Keywords: sentiment level, sentiment shock, explanatory power, abnormal stock return, beta

Procedia PDF Downloads 105

Authors: Omar Abdillah, Mirna Adriani

Abstract:

In this work, we perform new approach for analyzing public sentiment towards the presidential candidate in the 2014 Indonesian election that expressed in Twitter. In this study we propose such procedure for analyzing sentiment over Indonesian political message by understanding the behavior of Indonesian society in sending message on Twitter. We took different approach from previous works by utilizing punctuation mark and Indonesian sentiment lexicon that completed with the new procedure in determining sentiment towards the candidates. Our experiment shows the performance that yields up to 83.31% of average precision. In brief, this work makes two contributions: first, this work is the preliminary study of sentiment analysis in the domain of political message that has not been addressed yet before. Second, we propose such method to conduct sentiment analysis by creating decision making procedure in which it is in line with the characteristic of Indonesian message on Twitter.

Keywords: unsupervised sentiment analysis, political message, lexicon based, user behavior understanding

Procedia PDF Downloads 447

Authors: Yuhong Niu, Na Li, Chunlin Jin, Hansheng Ding

Abstract:

The study reviewed all types of patents applied by Shanghai health system to analyze how patent development in China from the year of 1990 to 2012. The study used quantitative and comparative analysis to investigate the change and trends of patent numbers, patent types, patent claims, forward citations, patent life, patent transactions, etc. Results reflected an obviously increased numbers of invention patents, applications, and authorizations and short-life patents, but the ratio of invention patents represented an up and down change. Forward citations and transactions ratio always kept at a low level. The results meant that the protection of intellectual property in the Shanghai health sector had made great progress and lots of positive changes due to incentive policies by local government. However, the low-quality patents, at the same time, increased rapidly. Thus, in the future, it is suggested that the quality management should be strengthened, and invents should be estimated before patent application. It is also suggested that the incentives for intellectual property should be optimized to promote the comprehensive improvement of patent quantity and quality.

Keywords: patent claims, forward citations, patent life, patent transactions ratio

Procedia PDF Downloads 134

Authors: Balamurugan L.

Abstract:

Artificial intelligence (AI) and its applications in Intellectual Property Rights (IPR) has been significantly growing in recent years. In last couple of years, AI tools for Patent Research and Patent Analytics have been well-stabilized in terms of accuracy of references and representation of identified patent insights. However, AI tools for Patent Prosecution and Patent Litigation are still in the nascent stage and there may be a significant potential if such market is explored further. Our research is primarily focused on identifying potential whitespaces and schematic algorithms to automate the Patent Prosecution and Patent Litigation Process of the Intellectual Property. The schematic algorithms may assist leading AI tool developers, to explore such opportunities in the field of Intellectual Property. Our research is also focused on identification of pitfalls of the AI. For example, Information Security and its impact in IPR, and Potential remediations to sustain the IPR in the digital environment.

Keywords: artificial intelligence, patent analytics, patent drafting, patent litigation, patent prosecution, patent research

Procedia PDF Downloads 40

Authors: Adel Assiri, Ahmed Emam, Hmood Al-Dossari

Abstract:

Sentiment analysis (SA) has received growing attention in Arabic language research. However, few studies have yet to directly apply SA to Arabic due to lack of a publicly available dataset for this language. This paper partially bridges this gap due to its focus on one of the Arabic dialects which is the Saudi dialect. This paper presents annotated data set of 4700 for Saudi dialect sentiment analysis with (K= 0.807). Our next work is to extend this corpus and creation a large-scale lexicon for Saudi dialect from the corpus.

Keywords: Arabic, sentiment analysis, Twitter, annotation

Procedia PDF Downloads 591

Authors: Anuta Mukherjee, Saswati Mukherjee

Abstract:

Data is growing everyday. Social networking sites such as Twitter are becoming an integral part of our daily lives, contributing a large increase in the growth of data. It is a rich source especially for sentiment detection or mining since people often express honest opinion through tweets. However, although sentiment analysis is a well-researched topic in text, this analysis using Twitter data poses additional challenges since these are unstructured data with abbreviations and without a strict grammatical correctness. We have employed collision theory to achieve sentiment analysis in Twitter data. We have also incorporated discourse analysis in the collision theory based model to detect accurate sentiment from tweets. We have also used the retweet field to assign weights to certain tweets and obtained the overall weightage of a topic provided in the form of a query. Hadoop has been exploited for speed. Our experiments show effective results.

Keywords: sentiment analysis, twitter, collision theory, discourse analysis

Procedia PDF Downloads 499

Authors: Riona Dighe

Abstract:

This research attempts to identify an alternative approach to the measurement of Gross National Happiness (GNH). It uses artificial intelligence (AI), incorporating natural language processing (NLP) and sentiment analysis to measure GNH. We use ‘off the shelf’ NLP models responsible for the sentiment analysis of a sentence as a building block for this research. We constructed an algorithm using NLP models to derive a sentiment analysis score against sentences. This was then tested against a sample of 20 respondents to derive a sentiment analysis score. The scores generated resembled human responses. By utilising the MLP classifier, decision tree, linear model, and K-nearest neighbors, we were able to obtain a test accuracy of 89.97%, 54.63%, 52.13%, and 47.9%, respectively. This gave us the confidence to use the NLP models against sentences in websites to measure the GNH of a country.

Keywords: artificial intelligence, NLP, sentiment analysis, gross national happiness

Procedia PDF Downloads 68

Authors: Jehan Fahim M. Alsulami

Abstract:

Academic research about sentiment analysis in sentiment analysis has obtained significant advancement over recent years and is flourishing from the collection of knowledge provided by various academic disciplines. In the current study, the status and development trend of the field of sentiment analysis in social networks is evaluated through a bibliometric analysis of academic publications. In particular, the distributions of publications and citations, the distribution of subject, predominant journals, authors, countries are analyzed. The collaboration degree is applied to measure scientific connections from different aspects. Moreover, the keyword co-occurrence analysis is used to find out the major research topics and their evolutions throughout the time span. The area of sentiment analysis in social networks has gained growing attention in academia, with computer science and engineering as the top main research subjects. China and the USA provide the most to the area development. Authors prefer to collaborate more with those within the same nation. Among the research topics, newly risen topics such as COVID-19, customer satisfaction are discovered.

Keywords: bibliometric analysis, sentiment analysis, social networks, social media

Procedia PDF Downloads 177

Authors: Salima Behdenna, Fatiha Barigou, Ghalem Belalem

Abstract:

Due to the quick increase in the volume of Arabic opinions posted on various social media, Arabic sentiment analysis has become one of the most important areas of research. Compared to English, there is very little works on Arabic sentiment analysis, in particular aspect-based sentiment analysis (ABSA). In ABSA, aspect extraction is the most important task. In this paper, we propose a semantic aspect-based sentiment analysis approach for standard Arabic reviews to extract explicit aspect terms and identify the polarity of the extracted aspects. The proposed approach was evaluated using HAAD datasets. Experiments showed that the proposed approach achieved a good level of performance compared with baseline results. The F-measure was improved by 19% for the aspect term extraction tasks and 55% aspect term polarity task.

Keywords: sentiment analysis, opinion mining, Arabic, aspect level, opinion, polarity

Procedia PDF Downloads 134

Authors: E. Y. Igde, S. Aydogan, F. E. Boran, D. Akay

Abstract:

Patent data have an increasingly important role in economic growth, innovation, technical advantages and business strategies and even in countries competitions. Analyzing of patent data is crucial since patents cover large part of all technological information of the world. In this paper, we have used the linguistic summarization technique to prove the validity of the hypotheses related to patent data stated in the literature.

Keywords: data mining, fuzzy sets, linguistic summarization, patent data

Procedia PDF Downloads 244

Authors: Gehad S. Kaseb, Mona F. Ahmed

Abstract:

Arabic Sentiment Analysis (SA) is one of the most common research fields with many open areas. Few studies apply SA to Arabic dialects. This paper proposes different pre-processing steps and a modified methodology to improve the accuracy using normal Support Vector Machine (SVM) classification. The paper works on two datasets, Arabic Sentiment Tweets Dataset (ASTD) and Extended Arabic Tweets Sentiment Dataset (Extended-AATSD), which are publicly available for academic use. The results show that the classification accuracy approaches 86%.

Keywords: Arabic, classification, sentiment analysis, tweets

Procedia PDF Downloads 112

Authors: Kujhin Jeong, Sungjoo Lee

Abstract:

In the past, cross-licensing was made up of various application or commercial patents. Today, cross-licensing is restricted to essential patents, which has emphasized their importance significantly. Literature has shown that patent portfolio provides information for patent protection or strategy decision-making, but little empirical research has found strategic tool of essential patents. This paper will highlight four types of essential patent portfolio and analysis about each strategy in the field of LTE-A. Specifically we collected essential patents of mobile communication company through ETSI (European Telecommunication Standards Institute) and build-up portfolio activity, concentration, diversity, and quality. Using these portfolios, we can understand each company’s strategic character about the technology of LTE-A and comparison analysis of financial results. Essential patents portfolio displays a mobile communication company’s strategy and its strategy’s impact on the performance of a company.

Keywords: essential patent, portfolio, patent portfolio, essential patent portfolio

Procedia PDF Downloads 353

Authors: Hanan Alshaher, Jinsheng Xu

Abstract:

Large amounts of data on the web can provide valuable information. For example, product reviews help business owners measure customer satisfaction. Sentiment analysis classifies texts into two polarities: positive and negative. This paper examines movie reviews and tweets using a new term weighting scheme, called one-over-sigma (1/sigma), on benchmark datasets for sentiment classification. The proposed method aims to improve the performance of sentiment classification. The results show that 1/sigma is more accurate than the popular term weighting schemes. In order to verify if the entropy reflects the discriminating power of terms, we report a comparison of entropy values for different term weighting schemes.

Keywords: 1/sigma, natural language processing, sentiment analysis, term weighting scheme, text classification

Procedia PDF Downloads 177

Authors: Sergio V Davalos, Alison L. Watkins

Abstract:

User generated content (UGC) such as website post has data associated with it: time of the post, gender, location, type of device, and number of words. The text entered in user generated content (UGC) can provide a valuable dimension for analysis. In this research, each user post is treated as a collection of terms (words). In addition to the number of words per post, the frequency of each term is determined by post and by the sum of occurrences in all posts. This research focuses on one specific aspect of UGC: sentiment. Sentiment analysis (SA) was applied to the content (user posts) of two sets of political forums related to the US presidential elections for 2012 and 2016. Sentiment analysis results in deriving data from the text. This enables the subsequent application of data analytic methods. The SASA (SAIL/SAI Sentiment Analyzer) model was used for sentiment analysis. The application of SASA resulted with a sentiment score for each post. Based on the sentiment scores for the posts there are significant differences between the content and sentiment of the two sets for the 2012 and 2016 presidential election forums. In the 2012 forums, 38% of the forums started with positive sentiment and 16% with negative sentiment. In the 2016 forums, 29% started with positive sentiment and 15% with negative sentiment. There also were changes in sentiment over time. For both elections as the election got closer, the cumulative sentiment score became negative. The candidate who won each election was in the more posts than the losing candidates. In the case of Trump, there were more negative posts than Clinton’s highest number of posts which were positive. KNIME topic modeling was used to derive topics from the posts. There were also changes in topics and keyword emphasis over time. Initially, the political parties were the most referenced and as the election got closer the emphasis changed to the candidates. The performance of the SASA method proved to predict sentiment better than four other methods in Sentibench. The research resulted in deriving sentiment data from text. In combination with other data, the sentiment data provided insight and discovery about user sentiment in the US presidential elections for 2012 and 2016.

Keywords: sentiment analysis, text mining, user generated content, US presidential elections

Procedia PDF Downloads 157

Authors: Yunwei Zhang, Na Li, Yuhong Niu

Abstract:

Bone tissue engineering scaffold was regarded as an important clinical technology of curing bony defect. The patent protection of bone tissue engineering scaffold had been paid more attention and strengthened all over the world. This study analyzed the future development trends of international technologies in the field of bone tissue engineering scaffold and its patent protection. This study used the methods of data classification and classification indexing to analyze 2718 patents retrieved in the patent database. Results showed that the patents coming from United States had a competitive advantage over other countiries in the field of bone tissue engineering scaffold. The number of patent applications by a single company in U.S. was a quarter of that of the world. However, the capability of R&D in China was obviously weaker than global level, patents mainly coming from universities and scientific research institutions. Moreover, it would be predicted that synthetic organic materials as new materials would be gradually replaced by composite materials. The patent technology protections of composite materials would be more strengthened in the future.

Keywords: bone tissue engineering, patent analysis, Scaffold material, patent protection

Procedia PDF Downloads 109

Authors: Yufen Qin

Abstract:

Text sentiment transfer is an important branch of text style transfer. The goal is to generate text with another sentiment attribute based on a text with a specific sentiment attribute while maintaining the content and semantic information unrelated to sentiment unchanged in the process. There are currently two main challenges in this field: no parallel corpus and text attribute entanglement. In response to the above problems, this paper proposed a novel solution: Mask-Prompt-Rerank. Use the method of masking the sentiment words and then using prompt regeneration to transfer the sentence sentiment. Experiments on two sentiment benchmark datasets and one formality transfer benchmark dataset show that this approach makes the performance of small pre-trained language models comparable to that of the most advanced large models, while consuming two orders of magnitude less computing and memory.

Keywords: language model, natural language processing, prompt, text sentiment transfer

Procedia PDF Downloads 45

Authors: Tian Lan

Abstract:

Arbitrageurs participate in a variety of techniques in response to the existence of fluctuating sentiment, resulting in sparse sentiment exposures. This paper found that Commodity Trading Advisor (CTA) funds in the top decile rated by sentiment beta outperformed those in the bottom decile by 0.33% per month on a risk-adjusted basis, with the difference being larger among skilled managers. This paper also discovered that around ten percent of Commodity Trading Advisor (CTA) funds could accurately predict market sentiment, which has a positive correlation with fund sentiment beta and acts as a determinant in fund performance. Instead of betting against mispricing, this research demonstrates that a competent manager can achieve remarkable returns by forecasting and reacting to shifts in investor sentiment.

Keywords: investment sentiment, CTA fund, market timing, fund performance

Procedia PDF Downloads 51

Authors: Kisik Song, Kyuwoong Kim, Sungjoo Lee

Abstract:

Many global firms and corporations derive new technology and opportunity by identifying vacant technology from patent analysis. However, previous studies failed to focus on technologies that promised continuous growth in industrial fields. Most studies that derive new technology opportunities do not test practical effectiveness. Since previous studies depended on expert judgment, it became costly and time-consuming to evaluate new technologies based on patent analysis. Therefore, research suggests a quantitative and systematic approach to technology evaluation indicators by using patent data to and from customer communities. The first step involves collecting two types of data. The data is used to construct evaluation indicators and apply these indicators to the evaluation of new technologies. This type of data mining allows a new method of technology evaluation and better predictor of how new technologies are adopted.

Keywords: data mining, evaluating new technology, technology opportunity, patent analysis

Procedia PDF Downloads 339

Authors: Vishnu Goyal, Basant Agarwal

Abstract:

Sentiment analysis research provides methods for identifying the people’s opinion written in blogs, reviews, social networking websites etc. Sentiment analysis is to understand what opinion people have about any given entity, object or thing. Sentiment analysis research can be broadly categorised into three types of approaches i.e. semantic orientation, machine learning and lexicon based approaches. Feature selection methods improve the performance of the machine learning algorithms by eliminating the irrelevant features. Information gain feature selection method has been considered best method for sentiment analysis; however, it has the drawback of selection of threshold. Therefore, in this paper, we propose a hybrid feature selection methods comprising of information gain and proposed feature selection method. Initially, features are selected using Information Gain (IG) and further more noisy features are eliminated using the proposed feature selection method. Experimental results show the efficiency of the proposed feature selection methods.

Keywords: feature selection, sentiment analysis, hybrid feature selection

Procedia PDF Downloads 298

Authors: Yu Zhang, Pedro Desouza

Abstract:

Since big data has become substantially more accessible and manageable due to the development of powerful tools for dealing with unstructured data, people are eager to mine information from social media resources that could not be handled in the past. Sentiment analysis, as a novel branch of text mining, has in the last decade become increasingly important in marketing analysis, customer risk prediction and other fields. Scientists and researchers have undertaken significant work in creating and improving their sentiment models. In this paper, we present a concept of selecting appropriate classifiers based on the features and qualities of data sources by comparing the performances of five classifiers with three popular social media data sources: Twitter, Amazon Customer Reviews, and Movie Reviews. We introduced a couple of innovative models that outperform traditional sentiment classifiers for these data sources, and provide insights on how to further improve the predictive power of sentiment analysis. The modelling and testing work was done in R and Greenplum in-database analytic tools.

Keywords: sentiment analysis, social media, Twitter, Amazon, data mining, machine learning, text mining

Procedia PDF Downloads 322

Authors: Adrian-Gabriel Chifu, Sebastien Fournier

Abstract:

One of the challenges of natural language understanding is to deal with the subjectivity of sentences, which may express opinions and emotions that add layers of complexity and nuance. Sentiment analysis is a field that aims to extract and analyze these subjective elements from text, and it can be applied at different levels of granularity, such as document, paragraph, sentence, or aspect. Aspect-based sentiment analysis is a well-studied topic with many available data sets and models. However, there is no clear definition of what makes a sentence difficult for aspect-based sentiment analysis. In this paper, we explore this question by conducting an experiment with three data sets: ”Laptops”, ”Restaurants”, and ”MTSC” (Multi-Target-dependent Sentiment Classification), and a merged version of these three datasets. We study the impact of domain diversity and syntactic diversity on difficulty. We use a combination of classifiers to identify the most difficult sentences and analyze their characteristics. We employ two ways of defining sentence difficulty. The first one is binary and labels a sentence as difficult if the classifiers fail to correctly predict the sentiment polarity. The second one is a six-level scale based on how many of the top five best-performing classifiers can correctly predict the sentiment polarity. We also define 9 linguistic features that, combined, aim at estimating the difficulty at sentence level.

Keywords: sentiment analysis, difficulty, classification, machine learning

Procedia PDF Downloads 41