🔬 Synthetic Gamma Spectra Training Data Analysis
📊 Dataset Summary
| Total Samples |
10,000 |
| Unique Isotopes |
40 |
| Avg Isotopes per Sample |
2.41 |
| Duration Range |
60.0s - 300.0s |
| Mean Duration |
179.6s |
| Activity Range |
1.01 - 99.99 Bq |
| Detectors |
radiacode_103 |
1. Isotope Distribution
What this shows: The frequency of each isotope across all training samples.
Imbalanced distributions may lead to model bias towards common isotopes.
2. Sample Complexity
What this shows: Distribution of how many source isotopes are present per sample.
Mix of single and multi-isotope samples helps the model handle real-world complexity.
3. Temporal & Activity Analysis
What this shows: Distribution of measurement durations and source activities.
Varied durations simulate different counting scenarios.
4. Isotope Co-occurrence
What this shows: Which isotopes frequently appear together in training samples.
This helps understand potential confusion pairs and realistic combinations.
5. Sample Spectra Visualization
What this shows: Actual spectrum shapes from the training data.
Each peak corresponds to gamma emission lines from the source isotopes.
3D Time-Energy-Counts View
6. Isotope Database Overview
What this shows: The complete isotope database structure organized by category.
Click to explore the hierarchy.