viewof mlModelType = Inputs.select(
["Supervised - Classification", "Unsupervised - Anomaly Detection", "Time-Series - RUL Prediction"],
{
label: "ML Model Type:",
value: "Supervised - Classification"
}
)
viewof trainingDataSize = Inputs.range([10, 500], {
label: "Training Examples:",
step: 10,
value: 100
})
viewof failureRate = Inputs.range([0.01, 0.20], {
label: "Historical Failure Rate:",
step: 0.01,
value: 0.05,
format: x => `${(x * 100).toFixed(0)}%`
})
mlPerformance = {
const baseMetrics = {
"Supervised - Classification": {
accuracy: 0.85,
precision: 0.82,
recall: 0.88,
f1: 0.85,
trainingTime: "4-8 hours",
inference: "< 100ms",
dataRequirement: "Labeled failure examples"
},
"Unsupervised - Anomaly Detection": {
accuracy: 0.75,
precision: 0.70,
recall: 0.92,
f1: 0.79,
trainingTime: "1-2 hours",
inference: "< 50ms",
dataRequirement: "Normal operation only"
},
"Time-Series - RUL Prediction": {
accuracy: 0.80,
precision: 0.78,
recall: 0.85,
f1: 0.81,
trainingTime: "8-16 hours",
inference: "< 200ms",
dataRequirement: "Degradation curves to failure"
}
};
const metrics = baseMetrics[mlModelType];
// Adjust accuracy based on training data size
const dataFactor = Math.min(trainingDataSize / 200, 1.0);
const adjustedAccuracy = metrics.accuracy * (0.7 + 0.3 * dataFactor);
const adjustedPrecision = metrics.precision * (0.7 + 0.3 * dataFactor);
const adjustedRecall = metrics.recall * (0.75 + 0.25 * dataFactor);
const adjustedF1 = 2 * (adjustedPrecision * adjustedRecall) / (adjustedPrecision + adjustedRecall);
// Calculate expected false positives and false negatives
const totalAssets = 100;
const expectedFailures = totalAssets * failureRate;
const truePositives = expectedFailures * adjustedRecall;
const falseNegatives = expectedFailures - truePositives;
const falsePositives = (totalAssets - expectedFailures) * (1 - metrics.precision) * 0.3;
return {
accuracy: adjustedAccuracy,
precision: adjustedPrecision,
recall: adjustedRecall,
f1: adjustedF1,
trainingTime: metrics.trainingTime,
inference: metrics.inference,
dataRequirement: metrics.dataRequirement,
truePositives: truePositives.toFixed(1),
falseNegatives: falseNegatives.toFixed(1),
falsePositives: falsePositives.toFixed(1)
};
}
html`<div style="background: #f8f9fa; padding: 20px; border-radius: 8px; border-left: 4px solid #3498DB; margin: 20px 0;">
<h3 style="margin-top: 0; color: #2C3E50;">${mlModelType}</h3>
<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(200px, 1fr)); gap: 15px; margin: 20px 0;">
<div style="background: white; padding: 15px; border-radius: 6px;">
<div style="color: #7F8C8D; font-size: 0.85em;">Accuracy</div>
<div style="font-size: 1.8em; font-weight: bold; color: #16A085;">
${(mlPerformance.accuracy * 100).toFixed(1)}%
</div>
</div>
<div style="background: white; padding: 15px; border-radius: 6px;">
<div style="color: #7F8C8D; font-size: 0.85em;">Precision</div>
<div style="font-size: 1.8em; font-weight: bold; color: #16A085;">
${(mlPerformance.precision * 100).toFixed(1)}%
</div>
</div>
<div style="background: white; padding: 15px; border-radius: 6px;">
<div style="color: #7F8C8D; font-size: 0.85em;">Recall</div>
<div style="font-size: 1.8em; font-weight: bold; color: #16A085;">
${(mlPerformance.recall * 100).toFixed(1)}%
</div>
</div>
<div style="background: white; padding: 15px; border-radius: 6px;">
<div style="color: #7F8C8D; font-size: 0.85em;">F1 Score</div>
<div style="font-size: 1.8em; font-weight: bold; color: #16A085;">
${(mlPerformance.f1 * 100).toFixed(1)}%
</div>
</div>
</div>
<div style="background: white; padding: 15px; border-radius: 6px; margin-top: 15px;">
<h4 style="margin-top: 0; color: #2C3E50;">Expected Performance (100 assets, ${(failureRate * 100).toFixed(0)}% failure rate)</h4>
<div style="display: grid; grid-template-columns: 1fr 1fr 1fr; gap: 15px;">
<div>
<div style="color: #16A085; font-size: 0.85em;">True Positives</div>
<div style="font-size: 1.5em; font-weight: bold; color: #16A085;">${mlPerformance.truePositives}</div>
<div style="color: #7F8C8D; font-size: 0.75em;">Correctly predicted failures</div>
</div>
<div>
<div style="color: #E74C3C; font-size: 0.85em;">False Negatives</div>
<div style="font-size: 1.5em; font-weight: bold; color: #E74C3C;">${mlPerformance.falseNegatives}</div>
<div style="color: #7F8C8D; font-size: 0.75em;">Missed failures</div>
</div>
<div>
<div style="color: #E67E22; font-size: 0.85em;">False Positives</div>
<div style="font-size: 1.5em; font-weight: bold; color: #E67E22;">${mlPerformance.falsePositives}</div>
<div style="color: #7F8C8D; font-size: 0.75em;">Unnecessary alerts</div>
</div>
</div>
</div>
<div style="display: grid; grid-template-columns: 1fr 1fr; gap: 15px; margin-top: 15px;">
<div style="background: white; padding: 12px; border-radius: 6px;">
<div style="color: #7F8C8D; font-size: 0.85em;">Training Time</div>
<div style="font-weight: bold; color: #2C3E50;">${mlPerformance.trainingTime}</div>
</div>
<div style="background: white; padding: 12px; border-radius: 6px;">
<div style="color: #7F8C8D; font-size: 0.85em;">Inference Time</div>
<div style="font-weight: bold; color: #2C3E50;">${mlPerformance.inference}</div>
</div>
</div>
<div style="background: #E67E2210; padding: 12px; border-radius: 6px; margin-top: 15px;">
<div style="color: #7F8C8D; font-size: 0.85em; margin-bottom: 5px;">Data Requirement</div>
<div style="color: #2C3E50; font-weight: 500;">${mlPerformance.dataRequirement}</div>
</div>
</div>`Time-Series Forecasting
Time-Series Forecasting
animation
iiot
maintenance
Try Change ML Model Type: and Training Examples: to compare performance. Then press Calculate expected performance.
Observe Watch accuracy, precision, recall, and F1 score as Training Examples: changes.
Explain Describe how Historical Failure Rate: changes expected failures and missed failures after you press Calculate expected performance.
Technical boundaries. Performance uses fixed base metrics with training data adjustments capped at 200 examples. Expected failures and alerts use 100 assets.
Time-Series Forecasting
Approach: Predicts remaining useful life (RUL) based on degradation trends.
Algorithms:
- LSTM neural networks
- Prophet (trend + seasonality)
- Gaussian Process Regression
Output: “How many hours/days until failure?” (RUL estimate with confidence interval)
Key metrics:
- Mean Absolute Error (MAE)
- Root Mean Square Error (RMSE)
- Percentage within 10%/20% tolerance