New Avenues in Opinion Mining and Sentiment Analysis
Summary (2 min read)
Common Sentiment Analysis Tasks
- The basic task of opinion mining is polarity classification.
- Reviews such as "thumbs up" versus "thumbs down," or "like" versus "dislike" are examples of polarity classification.
- Distinguishing between subjective and objective text helps classify the sentiment.
- Several studies suggest that managing these two tasks jointly might benefit overall performance.
- In such instances, it's important to identify topics and separate the opinions associated with each topic.
Evolution of Opinion Mining
- Currently, opinion mining and sentiment analysis rely on vector extraction to represent the most salient and important text features.
- The authors can use this vector to classify the most relevant features.
- Two commonly used features are term frequency and presence.
- Some methods also rely on the distance between terms.
- Also, selected phrases chosen by pre-specified POS patterns, usually including an adjective or adverb, help detect sentiments.
From heuristics to Discourse Structure
- In some unsupervised learning approaches, a sentiment lexicon is generated and later used to determine the text unit's degree of positivity or subjectivity.
- In their work on polarity classification, Vasileios Hatzivassiloglou and Kathleen Mc Keown discuss how two classes of interest represent opposites.
- Gradually the algorithm redefines the label with more specificity: first to a specific review corpus, then specific to a product feature, and finally to a label specific to the context in which the word occurs.
- Then they combined the prediction with individual-aspect classifiers, which would minimize loss function.
- 1 Modeling discourse structure, such as twists and turns in a document, leads to more effective sentiment labeling.
From coarse-to Fine-Grained analysis
- The authors see opinion mining and sentiment analysis research evolving in both technique sophistication and analysis depth.
- These documents were mainly supervised, manually labeled samples, such as movie or product reviews explicitly indicating an overall positive or negative opinion.
- In another study, Peter Turney classified items based on fixed, syntactic phrases used for expressing opinions.
- To correct that, other researchers adopted an aspect-level approach, wherein an opinion consists of targets and the sentiments associated with them. [15] [16] [17].
- The sentiment about the iPhone 5's screen size is positive, but the sentiment about its battery life is negative.
From Keywords to concepts
- The authors can study the evolution of sentiment analysis research by the analytical tokens, or building blocks, and the implicit information associated with those tokens.
- The authors can group the existing approaches into four main categories: keyword spotting, lexical affinity, statistical methods, and concept-based techniques.
Keyword spotting.
- Keyword spotting's accessibility and economy make it popular.
- Also, keyword spotting relies on the presence of obvious affect words that are only surface features of the prose.
- This approach not only detects obvious affect words, it also assigns arbitrary words a probable "affinity" to particular emotions.
- Lexical affinity might assign the word "accident" a 75-percent probability of indicating a negative affect, as in "car accident" or "hurt by accident.".
- This approach usually trains probability from linguistic corpora. [21] [22] [23].
Statistical methods.
- This approach, which includes Bayesian inference and support vector machines, is popular for affect text classification.
- As a result, statistical text classifiers only work well when they receive sufficiently large text input.
- This helps the system grasp the conceptual and affective information associated with natural language opinions.
- By relying on large semantic knowledge bases, such approaches step away from blindly using keywords and word cooccurrence counts, and instead rely on the implicit meaning/features associated with natural language concepts.
- Concept-based approaches can analyze multi-word expressions that don't explicitly convey emotion, but are related to concepts that do.
Multimodal Sentiment Analysis
- New sources of opinion mining and sentiment analysis abound.
- For a rough idea of the amount of material, consider that You-Tube users upload two days' worth of video material to its website every minute.
- Affect analysis, a related field, addresses the use of linguistic, acoustic, and video information.
- Again, the researchers relied on transcripts to analyze the text and not the actual spoken word.
- We'll then need to investigate analyses of robustness against disturbances in individual (or all) modalities alongside audiovisual confidence estimation.
Discussion
- Gradually, sentiment analysis research is distinguishing itself as a separate field, falling between NLP and natural language understanding.
- 32 Collecting opinions on the Web will still require processing at the content/ syntactic level, filtering out unopinionated user-generated content (subjectivity detection) and evaluating the trustworthiness of the opinion and its source.
- To make these comparisons, researchers must construct comprehensive common-knowledge bases to spot features and commonsense bases to detect polarity.
- Collective knowledge has spread throughout the Web, particularly in areas related to everyday life, such as commerce, tourism, education, and health.
- Opinion mining and sentiment analysis are inextricably bound to the affective sciences that attempt to understand human emotions.
Did you find this useful? Give us your feedback
Citations
2,152 citations
Cites background from "New Avenues in Opinion Mining and S..."
...[9], Feldman [10] and Montoyo and Martı́nez-Barco [11] have given short surveys illustrating the new trends in SA....
[...]
1,153 citations
1,139 citations
Cites background from "New Avenues in Opinion Mining and S..."
..., entity recognition, coreference resolution, negation handling) of Natural Language Processing (Liu, 2012) and as Cambria et al. (2013) mention “it requires a deep understanding of the explicit and implicit, regular and irregular, and syntactic and semantic language rules”....
[...]
...…(SA) touches every aspect (e.g., entity recognition, coreference resolution, negation handling) of Natural Language Processing (Liu, 2012) and as Cambria et al. (2013) mention “it requires a deep understanding of the explicit and implicit, regular and irregular, and syntactic and semantic…...
[...]
1,011 citations
Cites background from "New Avenues in Opinion Mining and S..."
...[15] pointed out complexities involved in SA with respect to current demand along with possible future research directions....
[...]
579 citations
Cites methods from "New Avenues in Opinion Mining and S..."
...When considering the future of aspect-level sentiment analysis, we foresee a move from traditional word-based approaches, towards semantically rich concept-centric aspect-level sentiment analysis [95]....
[...]
References
7,452 citations
7,330 citations
6,980 citations
6,792 citations
6,626 citations
Related Papers (5)
Frequently Asked Questions (5)
Q2. What is the common way to perform sentiment analysis?
a system performs sentiment analysis over on-topic documents— using, for example, the results of a topic-based search engine.
Q3. How did Ellen Riloff and Janyce Weibe reduce text-analysis gran?
In another work, Ellen Riloff and Janyce Weibe reduced text-analysis granularity to the sentence level by using the presence of opinion-bearing lexical items (single words or ngrams) to detect subjective sentences.
Q4. What are the two fields of opinion mining and sentiment analysis?
Both fields use data mining and natural language processing (NLP) techniques to discover, retrieve, and distill information and opinions from the World Wide Web’s vast textual information.
Q5. What are the main challenges of opinion mining?
If the text doesn’t contain strong opinions or covers more than one issue or item, new challenges arise, such as subjectivity detection and opinion-target identification.