Sentiment Analysis Overview

Neuredge's sentiment analysis capability helps you understand emotions, opinions, and attitudes expressed in text. Our models provide detailed sentiment insights that can be used for brand monitoring, customer feedback analysis, and social media intelligence.

Features

Core Capabilities

• Multi-level Analysis: Beyond just positive/negative classifications
• Emotion Detection: Identify specific emotions in text
• Aspect-based Analysis: Sentiment for specific aspects or topics
• Confidence Scores: Reliability metrics for predictions

Integration with Other Services

• Combines with summarization for efficient feedback processing
• Works with translation for multi-language sentiment analysis
• Pairs with embeddings for semantic sentiment search

Getting Started

Basic Usage

import { NeuredgeClient } from '@neuredge/sdk';

const client = new NeuredgeClient({
  apiKey: 'your-api-key'
});

// Simple sentiment analysis
const sentiment = await client.text.analyzeSentiment(
  'I really love this product! The features are amazing.'
);

// With more options
const detailedSentiment = await client.text.analyzeSentiment(
  'The interface is great but customer service needs improvement.',
  {
    includeEmotions: true,
    extractAspects: true,
    confidenceScores: true
  }
);

Python Example

from neuredge_sdk import NeuredgeClient

client = NeuredgeClient("YOUR_API_KEY")

# Basic sentiment analysis
sentiment = client.text.analyze_sentiment(
    "I really love this product! The features are amazing."
)

# Advanced options
detailed_sentiment = client.text.analyze_sentiment(
    "The interface is great but customer service needs improvement.",
    include_emotions=True,
    extract_aspects=True,
    confidence_scores=True
)

Advanced Features

Batch Analysis

const texts = [
  'This product is amazing!',
  'The service was terrible.',
  'Not sure how I feel about this.'
];

const batchSentiments = await client.text.analyzeSentimentBatch(
  texts,
  {
    includeEmotions: true
  }
);

Aspect-Based Analysis

const aspectSentiment = await client.text.analyzeSentiment(
  'The food was delicious but the service was slow.',
  {
    extractAspects: true,
    aspects: ['food', 'service', 'ambiance']
  }
);

Emotion Detection

const emotions = await client.text.analyzeEmotions(
  'I am so excited about this new feature!',
  {
    includeIntensity: true,
    emotionSet: 'extended'
  }
);

Best Practices

Optimizing Results

1. Text Preparation

   • Clean and normalize input text
   • Consider context when needed
   • Group related content for batch analysis

2. Analysis Configuration

   • Choose appropriate detail level
   • Set relevant aspects for tracking
   • Use confidence thresholds

3. Error Handling

try {
  const sentiment = await client.text.analyzeSentiment(
    'Your text here',
    { includeEmotions: true }
  );
} catch (error) {
  if (error.code === 'invalid_request') {
    console.log('Invalid request parameters');
  } else if (error.code === 'text_too_long') {
    console.log('Text exceeds maximum length');
  }
}

Use Cases

Customer Feedback Analysis

// Analyze customer review
const feedbackAnalysis = await client.text.analyzeSentiment(
  reviewText,
  {
    extractAspects: true,
    aspects: ['product', 'service', 'price', 'quality'],
    includeEmotions: true
  }
);

Social Media Monitoring

// Analyze social media post
const socialAnalysis = await client.text.analyzeSentiment(
  postText,
  {
    includeEmotions: true,
    extractHashtags: true,
    trackTrends: true
  }
);

Brand Monitoring

// Monitor brand mentions
const brandSentiment = await client.text.analyzeSentiment(
  brandMention,
  {
    aspectBasedAnalysis: true,
    brandAttributes: ['reputation', 'quality', 'innovation'],
    timeTracking: true
  }
);

Integration Examples

Feedback Processing Pipeline

async function processFeedback(feedback: string) {
  // Analyze sentiment
  const sentiment = await client.text.analyzeSentiment(
    feedback,
    {
      extractAspects: true,
      includeEmotions: true
    }
  );
  
  // Generate summary
  const summary = await client.text.summarize(
    feedback,
    { maxLength: 100 }
  );
  
  // Translate if needed
  const translation = await client.text.translate(
    summary,
    { targetLanguage: 'es' }
  );
  
  return {
    sentiment,
    summary,
    translation
  };
}

Real-time Sentiment Monitoring

async function monitorLiveFeedback(feedbackStream: string[]) {
  const results = [];
  
  for (const feedback of feedbackStream) {
    // Analyze sentiment
    const sentiment = await client.text.analyzeSentiment(
      feedback,
      { includeEmotions: true }
    );
    
    // Generate embedding for clustering
    const embedding = await client.embeddings.create({
      text: feedback,
      model: 'embedding-v1'
    });
    
    results.push({
      text: feedback,
      sentiment,
      embedding: embedding.vector
    });
  }
  
  return results;
}

Output Format

Basic Sentiment

{
  "sentiment": "positive",
  "score": 0.85,
  "confidence": 0.92
}

Detailed Analysis

{
  "overall": {
    "sentiment": "mixed",
    "score": 0.6,
    "confidence": 0.88
  },
  "aspects": {
    "product": {
      "sentiment": "positive",
      "score": 0.9
    },
    "service": {
      "sentiment": "negative",
      "score": 0.3
    }
  },
  "emotions": {
    "joy": 0.6,
    "frustration": 0.3,
    "satisfaction": 0.7
  }
}