Objective

To explore Aspect-Based Sentiment Analysis (ABSA) as a more granular approach to understanding sentiment beyond traditional polarity-based methods, helping businesses gain deeper insights into customer feedback, refine product offerings, enhance customer experience, and drive data-informed decision-making.

What is Aspect Based Sentiment Analysis

ABSA takes traditional sentiment analysis to the next level by evaluating sentiments at an attribute level. For example, in a review of a product, ABSA can separate opinions about the product’s quality, price, and customer service, identifying individual sentiments for each.

Tools & Libraries for Sentiment Analysis

Overall Sentiment Analysis Tools:

• • TextBlob: A simple library for rule-based sentiment scoring.

• • VADER: Ideal for short, informal text like social media posts.

• • Google NLP: Cloud-based API for sentiment scoring with robust support for large datasets.

As you can see from the above results, the Vader, TextBlob, and Google NLP models do not provide aspect-based sentiment scores for the given text. And hence we can try some specialized models, such as below, for a more accurate and meaningful aspect-based sentiment analysis.

Aspect Based Sentiment Analysis Tools:

HuggingFace Models: Pre-trained transformers for state-of-the-art sentiment scoring.

• • j-hartmann/sentiment-roberta-large-english-3-classes: Excellent for three-class sentiment analysis.

• • facebook/bart-large-mnli: Multi-task model for classification and aspect detection.

• • siebert/sentiment-roberta-large-english: Optimized for English language sentiment tasks.

• • yangheng/deberta-v3-base-absa-v1.1: Tailored for ABSA with state-of-the-art accuracy.

• • BertForSequenceClassification: Highly customizable for aspect sentiment fine-tuning.

Text Summarization for Long Reviews

Long texts often exceed token limits of many models (typically 512 tokens). Using summarization before performing ABSA can help reduce input size while retaining key information.

Tools include:

• • facebook/bart-large-cnn: An encoder-decoder model optimized for summarization tasks.

• • T5ForConditionalGeneration: A flexible transformer model supporting various NLP tasks, including summarization.

• • FalconsAI/text_summarization: A model generating context-aware summaries using advanced NLP techniques.

Preprocessing Methods for ABSA

Preprocessing is crucial to address noise, grammar issues, and formatting inconsistencies.

Tools include:

• • SentenceTransformer (all-MiniLM-L6-v2): Calculates semantic similarity between sentences for better aspect identification.

• • DeepMultilingualPunctuation: Fixes grammar and punctuation issues in text.

• • SpaCy/NLTK: For tokenization, stemming, lemmatization, and cleaning.

Challenges in ABSA

Handling Long Texts

Challenge: Many models limit input to 512 tokens, requiring chunking or summarization.

Solution: Summarize reviews using models like BART or split text into smaller chunks for analysis.

Poor Grammar and Punctuation

Challenge: Improper grammar or punctuation can confuse models and reduce accuracy.

Solution: Use punctuation models (e.g., DeepMultilingualPunctuation) and regular expressions to clean text.

Special Characters

Solution: Clean and normalize text using regex and preprocessing libraries like NLTK.

Challenge: Special characters and symbols disrupt model comprehension.

Step-by-Step ABSA Process

Preprocessing

• • Fix punctuation and grammar issues.

• • Remove special characters.

• • Tokenize, lemmatize, and clean the text.

Sentiment Analysis

• • Perform overall sentiment analysis using VADER or TextBlob to get a high-level view.

• • Identify specific aspects using Hugging Face models (e.g., deberta-v3-base-absa-v1.1).

Summarization (if required)

For long texts, summarize using facebook/bart-large-cnn or T5.

Aggregation of Results

Combine aspect scores to provide a holistic sentiment overview.

Example: ABSA in Action

Original Text (Before Preprocessing):

“3M is an amazingly innovative company. As a scientist, working at 3M is like being a kid in a toy store. Unfortunately, management prioritizes shareholders over employees, with major layoffs during the pandemic despite high dividends for investors.”

Challenges Identified:

• • Grammar issues and lack of punctuation.
    Text exceeds token limit. 

After Preprocessing (Corrected and Summarized):

“3M is innovative but prioritizes shareholders over employees. Layoffs during the pandemic despite high dividends highlight leadership issues. A flatter structure and better leadership could improve morale.”

Addressing Challenges with Long Texts

Method 1: Chunk-Based Sentiment Analysis

Split text into chunks and analyze sentiment for each.

Aggregate chunk results for an overall sentiment score.
Pros: Captures sentence-level nuances.
Cons: Time-intensive and less accurate for overall analysis.

Method 2: Summarization First

Summarize long texts to within 500 words.

Perform ABSA on the summarized content.
Pros: Faster and more accurate for overall sentiment.
Cons: Risk of losing subtle details.

Conclusions

• • ABSA enables businesses to derive actionable insights by analyzing specific attributes like leadership, innovation, and employee treatment.

• • Preprocessing is critical for handling noisy or unstructured text.

• • Hugging Face models like deberta-v3-base-absa-v1.1 are ideal for achieving state-of-the-art accuracy in ABSA tasks.

• • Addressing challenges like long texts and poor grammar improves the overall performance of sentiment models.