Class AgentAssistanceOptions

Namespace: AiDotNet.Models

Assembly: AiDotNet.dll

Configures which types of assistance the AI agent provides during model building.

public class AgentAssistanceOptions

Inheritance: object

AgentAssistanceOptions

Inherited Members: object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

This class allows fine-grained control over which aspects of model building the AI agent assists with. Each property enables or disables a specific capability, allowing you to balance between AI assistance and manual control based on your needs. The class provides three predefined configurations (Default, Minimal, and Comprehensive) as well as the ability to create custom configurations using the fluent builder pattern.

For Beginners: This class lets you choose exactly what kind of help you want from the AI agent.

Think of it like choosing which tasks to delegate to an assistant:

Data Analysis: Agent examines your data and reports on its characteristics
Model Selection: Agent recommends which machine learning algorithm to use
Hyperparameter Tuning: Agent suggests optimal settings for your chosen model
Feature Analysis: Agent identifies which input variables are most important
Meta-Learning Advice: Agent provides guidance on advanced few-shot learning setups

Common usage patterns:

Default: Good starting point, enables basic data analysis and model selection
Minimal: Only helps choose a model, gives you full control over everything else
Comprehensive: Maximum assistance, agent helps with every aspect
Custom: Use the Create() builder to enable only specific features you want

For example, if you're new to machine learning, you might use Comprehensive to get maximum help. If you're experienced but want validation on model choice, you might use Minimal. If you want help with everything except hyperparameters (which you want to tune yourself), you'd create a custom configuration.

Properties

Comprehensive

Gets a predefined configuration with all assistance features enabled.

public static AgentAssistanceOptions Comprehensive { get; }

Property Value

AgentAssistanceOptions: An AgentAssistanceOptions instance with maximum assistance suitable for getting the best results.

Remarks

For Beginners: This preset turns on every available AI assistance feature for maximum help.

What's enabled:

✓ Data analysis (understands your data)
✓ Model selection (picks the best algorithm)
✓ Hyperparameter tuning (optimizes settings)
✓ Feature analysis (identifies important variables)
✓ Meta-learning advice (if using meta-learning)

Use this when:

You want the AI to help with every decision
You're working on an important project and want maximum performance
You want to learn what the AI recommends for each aspect
You're new to ML and want comprehensive guidance

Note: This will make model building take longer because the agent does more analysis, but you'll get more insights and potentially better results.

Default

Gets a predefined configuration with data analysis and model selection enabled.

public static AgentAssistanceOptions Default { get; }

Property Value

AgentAssistanceOptions: An AgentAssistanceOptions instance with default settings suitable for most users.

Remarks

For Beginners: This is a balanced preset that enables the most commonly useful features.

What's enabled:

✓ Data analysis (helps you understand your data)
✓ Model selection (recommends which algorithm to use)
✗ Hyperparameter tuning (uses default settings)
✗ Feature analysis (keeps all features)
✗ Meta-learning advice (not needed for standard problems)

Use this when: You're getting started and want helpful guidance without overwhelming detail. It's a good balance between AI assistance and keeping things simple.

EnableDataAnalysis

Gets or sets a value indicating whether the agent should analyze data characteristics.

public bool EnableDataAnalysis { get; set; }

Property Value

bool: True to enable data analysis; false to disable. Default is true.

Remarks

When enabled, the agent examines your dataset and provides insights about its characteristics, such as distribution patterns, potential outliers, missing values, and data quality issues. This analysis helps identify potential problems before training begins.

For Beginners: This tells the agent to look at your data and report what it finds.

What the agent checks:

How many samples and features you have
Whether features are on similar scales
If there are outliers (unusual values)
Distribution patterns (normal, skewed, etc.)
Potential data quality issues

This is useful because:

You might discover problems before wasting time training
The agent can warn about imbalanced data or scaling issues
You learn about your data's characteristics

Example: The agent might report "Your dataset has 1000 samples with 20 features. Feature 'income' has much larger values than other features and should be normalized. Detected 5 potential outliers in feature 'age'."

EnableFeatureAnalysis

Gets or sets a value indicating whether the agent should analyze feature importance.

public bool EnableFeatureAnalysis { get; set; }

Property Value

bool: True to enable feature analysis; false to disable. Default is false.

Remarks

When enabled, the agent evaluates which input features are most important for making predictions and may suggest feature selection strategies. This can improve model performance, reduce training time, and increase interpretability by focusing on the most relevant features.

For Beginners: Features are the input variables you use to make predictions.

What the agent does:

Identifies which features actually matter for predictions
Detects features that add no value (can be removed)
Finds redundant features (similar information)
Suggests feature engineering opportunities

Why this matters:

More features isn't always better
Irrelevant features add noise and slow down training
Understanding which features matter helps you understand your problem
You might be able to collect less data if you know what's important

For example, when predicting house prices:

The agent might find that square footage and location are critical
The color of the front door probably doesn't matter
Number of bedrooms might be redundant with square footage

Why it's disabled by default:

Feature analysis requires additional computation
Not all problems benefit from feature selection
You might want to keep all features even if some are less important

Enable this when: You have many features and suspect some aren't useful, or you want to understand what drives your predictions.

EnableHyperparameterTuning

Gets or sets a value indicating whether the agent should recommend hyperparameter values.

public bool EnableHyperparameterTuning { get; set; }

Property Value

bool: True to enable hyperparameter tuning assistance; false to disable. Default is false.

Remarks

When enabled, the agent suggests optimal hyperparameter values for your chosen model based on dataset characteristics and best practices. Hyperparameters control how the model learns and can significantly impact performance. This feature is disabled by default as it requires more computational analysis and may not be necessary for all use cases.

For Beginners: Hyperparameters are settings that control how your model learns.

Think of hyperparameters like knobs on a radio:

Learning Rate: How quickly the model adjusts (too fast = unstable, too slow = takes forever)
Regularization: How much to penalize complex models (prevents overfitting)
Tree Depth: For tree-based models, how detailed the decisions can get
Number of Neurons: For neural networks, how many processing units to use

Why enable this:

Default hyperparameters often aren't optimal for your specific data
Bad hyperparameters can dramatically hurt performance
Finding good values usually requires trial and error
The agent can suggest good starting points based on your data

Why it's disabled by default:

It requires more time and analysis
Many datasets work fine with default settings
Advanced users often want to tune these themselves

Enable this when: You want to squeeze out maximum performance or when default settings aren't working well.

EnableMetaLearningAdvice

Gets or sets a value indicating whether the agent should provide meta-learning configuration advice.

public bool EnableMetaLearningAdvice { get; set; }

Property Value

bool: True to enable meta-learning assistance; false to disable. Default is false.

Remarks

When enabled, the agent provides guidance on meta-learning configurations such as task distribution, adaptation strategies, and inner/outer loop settings. This is an advanced feature primarily useful when working with few-shot learning scenarios where models must quickly adapt to new tasks.

For Beginners: Meta-learning is an advanced technique where models "learn to learn."

What is meta-learning:

Regular learning: Train a model on lots of data for one specific task
Meta-learning: Train a model on many different tasks so it can quickly learn new tasks with minimal data
It's like teaching someone how to learn quickly rather than teaching them one specific thing

When you'd use this:

You have many related tasks (e.g., image classification for different product categories)
Each task has only a few examples (few-shot learning)
You need models that can adapt quickly to new scenarios
You're working on personalization or rapid adaptation problems

What the agent helps with:

How to structure your tasks for meta-learning
Which meta-learning algorithm to use (MAML, Reptile, etc.)
How many adaptation steps to use
Balance between learning speed and final performance

Why it's disabled by default:

This is an advanced technique not needed for most projects
Regular supervised learning is sufficient for most use cases
Requires understanding of meta-learning concepts

Only enable this if: You're specifically working on few-shot learning or need models that rapidly adapt to new tasks.

EnableModelSelection

Gets or sets a value indicating whether the agent should suggest model types.

public bool EnableModelSelection { get; set; }

Property Value

bool: True to enable model selection assistance; false to disable. Default is true.

Remarks

When enabled and no model has been explicitly configured, the agent analyzes your data characteristics and recommends which machine learning algorithm would be most appropriate. The recommendation is based on dataset size, feature count, problem complexity, and expected performance trade-offs.

For Beginners: This lets the agent recommend which type of machine learning model to use.

Why this matters:

There are many different ML algorithms (linear regression, random forests, neural networks, etc.)
Each works best for different types of problems
Choosing the wrong model can give poor results
The agent considers your data's characteristics to make a smart recommendation

What the agent considers:

How much data you have (small datasets → simpler models)
Number of features (many features → models that handle high dimensions well)
Data patterns (linear relationships → linear models, complex patterns → tree-based or neural networks)
Your need for interpretability vs. accuracy

For example, with a small dataset of house prices based on size and location, the agent might recommend linear regression. For a large dataset with complex relationships, it might suggest a random forest or gradient boosting model.

Minimal

Gets a predefined configuration with only model selection enabled.

public static AgentAssistanceOptions Minimal { get; }

Property Value

AgentAssistanceOptions: An AgentAssistanceOptions instance with minimal assistance suitable for experienced users.

Remarks

For Beginners: This preset gives you just one piece of help: choosing which model to use.

What's enabled:

✗ Data analysis (you'll analyze data yourself)
✓ Model selection (agent recommends an algorithm)
✗ Hyperparameter tuning (you'll tune manually)
✗ Feature analysis (you'll handle features yourself)
✗ Meta-learning advice (not needed)

Use this when: You're experienced with ML and only want validation on which model type to use. Everything else you'll handle yourself. This gives you maximum control with one helpful suggestion.

Methods

Create()

Creates a new fluent builder for custom agent assistance configuration.

public static AgentAssistanceOptionsBuilder Create()

Returns

AgentAssistanceOptionsBuilder: A new AgentAssistanceOptionsBuilder instance for fluent configuration.

Remarks

For Beginners: Use this to create a custom configuration by picking exactly which features you want.

Example of creating a custom configuration:

var options = AgentAssistanceOptions.Create()
    .EnableDataAnalysis()
    .EnableModelSelection()
    .EnableFeatureAnalysis()
    .Build();

This gives you precise control - maybe you want data analysis and model selection but not hyperparameter tuning. The fluent builder makes it easy to enable exactly what you want.

Table of Contents

Class AgentAssistanceOptions

Remarks

Properties

Comprehensive

Property Value

Remarks

Default

Property Value

Remarks

EnableDataAnalysis

Property Value

Remarks

EnableFeatureAnalysis

Property Value

Remarks

EnableHyperparameterTuning

Property Value

Remarks

EnableMetaLearningAdvice

Property Value

Remarks

EnableModelSelection

Property Value

Remarks

Minimal

Property Value

Remarks

Methods

Create()

Returns

Remarks