Skill ML.NET Anomaly Detection

Detect anomalies in a single sensor tag using ML.NET SSA (Singular Spectrum Analysis). Supports: Spike Detection for sudden outliers. The AI generates the full C# Script Class, connects it to a live tag, creates output tags, and configures model persistence.

AI Integration → Platform Skills Library → Skill ML.NET → Skill ML.NET Anomaly Detection

When to Use This Skill

User chose Anomaly Detection — SSA Spike in the ML.NET router skill (Step 0)
Monitoring a single sensor for outliers, spikes, drift, or regime changes
User goals: predictive maintenance (single sensor), detect sensor failures, detect gradual process shift

Prerequisites

Solution open with the input tag already created and receiving data (live or simulated)
Solution target platform set to Multiplatform (ML.NET requires .NET 8+)
The user has confirmed: (1) which algorithm variant (Spike), (2) which tag to monitor, (3) output folder path

MCP Tools and Tables

Category	Items
Tools	`get_table_schema`, `write_objects`, `get_objects`
Tables	`UnsTags`, `ScriptsClasses`, `ScriptsExpressions`, `ScriptsTasks`

Step 1: Create Output Tags

Create tags to receive the model's predictions. Place them under a /ML/ subfolder relative to the input tag's asset path.

get_table_schema('UnsTags')

{
  "table_type": "UnsTags",
  "data": [
    { "Name": "<AssetPath>/ML/AnomalyScore", "DataType": "Double", "Description": "Anomaly score (0=normal, higher=anomalous)" },
    { "Name": "<AssetPath>/ML/IsAnomaly", "DataType": "Boolean", "Description": "True when anomaly detected" },
    { "Name": "<AssetPath>/ML/LastPrediction", "DataType": "DateTime", "Description": "Timestamp of last prediction" }
  ]
}

Replace <AssetPath> with the actual asset folder path (e.g., Plant/Reactor1).

Step 2: Create the Script Class

ML.NET Namespace Declaration

Script Classes using ML.NET require the NamespaceDeclarations field. ML.NET is pre-installed with FrameworX — no NuGet packages needed.

Critical: The field AddMLNetNamespaces does not exist and is silently ignored. Always use NamespaceDeclarations with a semicolon-separated string. Omitting it causes CS0246 / CS0234 compilation errors.

"NamespaceDeclarations": "Microsoft.ML;Microsoft.ML.Data;Microsoft.ML.Transforms;Microsoft.ML.Transforms.TimeSeries;Microsoft.ML.Transforms.Text;Microsoft.ML.Trainers;Microsoft.ML.TimeSeries"

Tag References Inside Script Classes

Always use the @Tag. prefix to read or write tag values:

// Read from a tag
double temp = @Tag.Plant/Reactor1/Temperature.Value;

// Write to a tag
@Tag.Plant/Reactor1/ML/AnomalyScore.Value = score;
@Tag.Plant/Reactor1/ML/IsAnomaly.Value = true;
@Tag.Plant/Reactor1/ML/LastPrediction.Value = DateTime.Now;

Important: ML.NET expects float but FrameworX tags use double. Always cast with (float) when feeding ML.NET and cast back to double when writing to tags.

SSA Spike Detection Pipeline

Use for sudden outliers, spikes, abnormal readings (e.g., pressure spikes, vibration bursts, temperature jumps).

var pipeline = mlContext.Transforms.DetectSpikeBySsa(
    outputColumnName: "Prediction",
    inputColumnName: nameof(SensorData.Value),
    confidence: 95.0,
    pvalueHistoryLength: 10,
    trainingWindowSize: 100,
    seasonalityWindowSize: 10);

Output: double[] Prediction with [0]=isAnomaly, [1]=score, [2]=pValue

Full Class Example — Anomaly Detection (Spike)

public class SensorData
{
    public float Value { get; set; }
}

public class SpikePrediction
{
    [VectorType(3)]
    public double[] Prediction { get; set; }
}

private static MLContext mlContext = new MLContext(seed: 0);
private static ITransformer model;
private static IDataView lastTrainingDataView;
private static bool modelTrained = false;
private static List<SensorData> trainingBuffer = new List<SensorData>();
private const int MinTrainingSize = 100;
private static readonly string ModelPath = Path.Combine(@Info.GetExecutionPath(), "<ClassName>.mlnet");

public int Predict(double inputValue)
{
    trainingBuffer.Add(new SensorData { Value = (float)inputValue });

    if (!modelTrained && trainingBuffer.Count >= MinTrainingSize)
        TrainModel();

    if (modelTrained)
        RunPrediction();

	return 1;
}

public void LoadModel()
{
    if (File.Exists(ModelPath))
    {
        model = mlContext.Model.Load(ModelPath, out _);
        modelTrained = true;
    }
}

private void TrainModel()
{
    lastTrainingDataView = mlContext.Data.LoadFromEnumerable(trainingBuffer);
    var pipeline = mlContext.Transforms.DetectSpikeBySsa(
        outputColumnName: "Prediction",
        inputColumnName: nameof(SensorData.Value),
        confidence: 95.0,
        pvalueHistoryLength: 10,
        trainingWindowSize: 100,
        seasonalityWindowSize: 10);

    model = pipeline.Fit(lastTrainingDataView);
    modelTrained = true;
    SaveModel();
}

private void SaveModel()
{
    mlContext.Model.Save(model, lastTrainingDataView.Schema, ModelPath);
}

private void RunPrediction()
{
    var dataView = mlContext.Data.LoadFromEnumerable(trainingBuffer);
    var transformed = model.Transform(dataView);
    var predictions = mlContext.Data.CreateEnumerable<SpikePrediction>(transformed, reuseRowObject: false).ToList();
    var latest = predictions.Last();

    @Tag.<AssetPath>/ML/IsAnomaly.Value = (int)latest.Prediction[0];
    @Tag.<AssetPath>/ML/AnomalyScore.Value = latest.Prediction[1];
    @Tag.<AssetPath>/ML/LastPrediction.Value = DateTime.Now;
}

Step 3: Write the Class via MCP

get_table_schema('ScriptsClasses')

{
  "table_type": "ScriptsClasses",
  "data": [
    {
      "Name": "<ClassName>",
      "Code": "CSharp",
      "Domain": "Server",
      "ClassContent": "Methods",
      "NamespaceDeclarations": "Microsoft.ML;Microsoft.ML.Data;Microsoft.ML.Transforms;Microsoft.ML.Transforms.TimeSeries;Microsoft.ML.Transforms.Text;Microsoft.ML.Trainers;Microsoft.ML.TimeSeries",
      "Contents": "<AI-generated C# code from full class example above>"
    }
  ]
}

Field names matter: the language field is Code (not Language), and the code body field is Contents (not Code). Using wrong field names results in silent data loss.

Step 4: Create the Trigger

Expression (OnChange) — triggers on each tag value change

get_table_schema('ScriptsExpressions')

{
  "table_type": "ScriptsExpressions",
  "data": [
    {
      "Name": "ML_Predict_<SensorName>",
      "ObjectName": "",
      "Expression": "@Script.Class.<ClassName>.Predict(@Tag.<AssetPath>.<Member>)",
      "Execution": "OnChange",
      "Trigger": ""
    }
  ]
}

ObjectName must be empty when the ML class writes prediction results to output tags internally (inside RunPrediction()). Setting a non-empty ObjectName on a void method call causes a type assignment error.

Use Trigger (not TriggerTag) — TriggerTag is not a valid field and is silently ignored, causing the expression to never fire. Trigger accepts the tag path without the Tag. prefix.

ServerStartup — always wire LoadModel

Read the existing ServerStartup task first (document object — read-modify-write), then append:

Script.Class.<ClassName>.LoadModel();

Step 5: Verify

Confirm Multiplatform — ML.NET requires .NET 8+. If the solution targets Windows (.NET 4.8), training will fail with a System.Math / CpuMath error. Instruct the user: “Before starting the runtime, please confirm your solution is set to Multiplatform: Solution → Settings → Target Platform = Multiplatform, then Product → Modify.”
Do NOT start the runtime automatically. Inform the user that all scripts are configured and they can start the runtime when ready.
Wait for training — the model needs MinTrainingSize data points (default 100) before predictions begin
Check output tags — verify LastPrediction timestamp is updating

Common Pitfalls

Mistake	Why It Happens	How to Avoid
Missing ML.NET namespaces	Used `AddMLNetNamespaces: true` (field does not exist)	Always set `NamespaceDeclarations` with the full semicolon-separated string
`CS0234` error in ScriptsTasks	Called `Script.Class.Name.Predict()` without `@` prefix	Always use `@Script.Class.Name.Predict()`
Non-empty `ObjectName` on void Predict()	Expression tries to assign void return to a tag	Leave `ObjectName` empty
Used `TriggerTag` field	Field does not exist — silently ignored	Use `Trigger` field
`CS0029` bool-to-int on Digital tag	Digital tags are `int`, not `bool`	Use ternary: `@Tag.Path.Value = (condition) ? 1 : 0;`
`CS0246` Script Class not found	Class compiles after Task that references it	Set `BuildOrder: "1"` on the ML Script Class
`System.Math` load error at training	Solution targets .NET 4.8	Switch to Multiplatform (.NET 8+)
Wrong data types	ML.NET expects `float`, tags are `double`	Cast with `(float)` for ML.NET, `(double)` for tags
Model lost on restart	SaveModel or LoadModel not wired	Always include both + wire LoadModel in ServerStartup

Page tree