Class SentimentAnalysis

Namespace: LMKit.TextAnalysis

Assembly: LM-Kit.NET.dll

Provides functionality for performing sentiment analysis on plain text, designed to assess and categorize emotional tone.

public sealed class SentimentAnalysis

Inheritance: object

SentimentAnalysis

Inherited Members: object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.ReferenceEquals(object, object)

object.ToString()

Examples

Example: Basic sentiment analysis

using LMKit.TextAnalysis;
using LMKit.Model;
using System;

// Initialize the language model (LM) using the specified model path
LM languageModel = LM.LoadFromModelID("lmkit-tasks:4b-preview");

// Create an instance of SentimentAnalysis
SentimentAnalysis sentimentAnalyzer = new SentimentAnalysis(languageModel);

// Analyze text sentiment
string text = "I absolutely love this product! It exceeded my expectations.";
SentimentCategory sentiment = sentimentAnalyzer.GetSentimentCategory(text);

Console.WriteLine($"Sentiment: {sentiment}");
// Output: "Sentiment: Positive"

Example: Batch analysis with confidence filtering

using LMKit.TextAnalysis;
using LMKit.Model;

LM model = LM.LoadFromModelID("lmkit-tasks:4b-preview");
SentimentAnalysis analyzer = new SentimentAnalysis(model);

string[] reviews = { "Great product!", "Terrible experience.", "It was okay." };
float confidenceThreshold = 0.8f;

foreach (var review in reviews)
{
    var sentiment = analyzer.GetSentimentCategory(review);
    if (analyzer.Confidence >= confidenceThreshold)
        Console.WriteLine($"[{sentiment}] {review}");
    else
        Console.WriteLine($"[UNCERTAIN] {review}");
}

Example: Including neutral sentiment

using LMKit.TextAnalysis;
using LMKit.Model;

LM model = LM.LoadFromModelID("lmkit-tasks:4b-preview");
SentimentAnalysis analyzer = new SentimentAnalysis(model);

// Enable neutral category
analyzer.NeutralSupport = true;

// Now factual statements can be classified as neutral
var sentiment = analyzer.GetSentimentCategory("The meeting is at 3 PM.");
Console.WriteLine($"Sentiment: {sentiment}"); // Output: Neutral

Remarks

The SentimentAnalysis class uses a language model to analyze textual content and determine its sentiment category. It supports both synchronous and asynchronous operations and allows for fine-tuning with custom training data.

Key Features

Classify text as positive, negative, or optionally neutral
Get confidence scores to evaluate classification reliability
Support for embedding-based classification for improved performance
Built-in training datasets for model fine-tuning
Multilingual support depending on the underlying model

Sentiment Categories

Category	Meaning
Positive	Indicates approval, satisfaction, enthusiasm, or favorable opinion
Negative	Indicates criticism, dissatisfaction, frustration, or unfavorable opinion
Neutral	Indicates factual, objective content with no strong emotional tone (requires NeutralSupport = true)

Classification Modes
By default, the analyzer uses completion-based classification. For certain models and high-volume scenarios, embedding-based classification (UseEmbeddingClassifier) can provide better performance.

Constructors

SentimentAnalysis(LM): Initializes a new instance of the SentimentAnalysis class.

Properties

Confidence: Gets the confidence score of the last sentiment analysis process, ranging from 0 to 1. A score closer to 1 indicates higher confidence in the categorization accuracy.

Model: Gets the LM instance associated with this SentimentAnalysis object.

NeutralSupport: Gets or sets a value indicating whether the neutral category support is enabled.

UseEmbeddingClassifier: Gets or sets a value indicating whether the classifier should utilize the embeddings strategy instead of completion.

Methods

GetSentimentCategory(string, CancellationToken): Analyzes the sentiment of the specified text and classifies it into a category defined in the SentimentAnalysis.SentimentCategory enumeration.

GetSentimentCategoryAsync(string, CancellationToken): Asynchronously analyzes the sentiment of the specified text and classifies it into a category defined in the SentimentAnalysis.SentimentCategory enumeration.