deep learning notebook, lots of fixes

README.md

@@ -91,7 +91,6 @@ Go to your working directory and run notebooks in order:
 
 Note:
 
-* Notebooks 01-03 are uploaded with all of their cells run, so that one can see the data analysis. Notebooks 04-06 are uploaded without having been run.
 * Notebook 03 generates hold difficulty tables
 * Notebook 04 generates feature matrix
 * Notebook 05 trains models
@@ -248,7 +247,7 @@ Here are some relationships between features and difficulty
 ![Correlation Heatmap](images/04_climb_features/feature_correlations.png)
 
 * higher angles allow for harder difficulties
-* distance between holds seems to correlate with difficulty
+* distance between holds seems to relate to difficulty
 * geometric and structural features capture non-trivial climbing patterns
 
 We have a full feature list in [`data/04_climb_features/feature_list.txt`](data/04_climb_features/feature_list.txt). Explanations are available in [`data/04_climb_features/feature_explanations.txt`](data/04_climb_features/feature_explanations.txt).
@@ -337,7 +336,6 @@ The model can still predict subgrades (e.g., V3 contains 6a and 6a+), but it is
 # Future Work
 
 * Unified grade prediction across boards
-* Combined board analysis
 * Test other models
 * Better spatial features
 * GUI to create climb and instantly tell you a predicted difficulty

data/06_deep_learning/neural_network_summary.txt

@@ -0,0 +1,33 @@
+
+### Neural Network Model Summary
+
+**Architecture:**
+- Input: 48 features
+- Hidden layers: [256, 128, 64]
+- Dropout rate: 0.2
+- Total parameters: 54,657
+
+**Training:**
+- Optimizer: Adam (lr=0.001)
+- Early stopping: 25 epochs patience
+- Best epoch: 109
+
+**Test Set Performance:**
+- MAE: 1.885
+- RMSE: 2.401
+- R²: 0.644
+- Accuracy within ±1 grade: 33.1%
+- Accuracy within ±2 grades: 60.8%
+- Exact grouped V-grade accuracy: 27.1%
+- Accuracy within ±1 V-grade: 69.4%
+- Accuracy within ±2 V-grades: 89.5%
+
+**Key Findings:**
+1. The neural network is competitive, but not clearly stronger than the best tree-based baseline.
+2. Fine-grained score prediction remains harder than grouped grade prediction.
+3. The grouped V-grade metrics show that the model captures broader difficulty bands more reliably than exact score labels.
+4. This makes the neural network useful as a comparison model, and potentially valuable in an ensemble.
+
+**Portfolio Interpretation:**
+This deep learning notebook extends the classical modelling pipeline by testing whether a neural architecture can improve prediction quality on engineered climbing features.
+The main result is not that deep learning wins outright, but that it provides a meaningful benchmark and helps clarify where model complexity does and does not add value.

notebooks/05_predictive_modelling.ipynb

@@ -363,6 +363,7 @@
     "    ax.set_title(f'{model_name}: Residual Plot', fontsize=14)\n",
     "\n",
     "    plt.tight_layout()\n",
+    "    plt.savefig(f'../images/05_predictive_modelling/{model_name.lower().replace(\" \", \"_\")}_predictions.png', dpi=150, bbox_inches='tight')\n",
     "    plt.show()\n",
     "\n",
     "\n",
@@ -394,6 +395,7 @@
     "    plt.suptitle('')  # Remove automatic title\n",
     "\n",
     "    plt.tight_layout()\n",
+    "    plt.savefig(f'../images/05_predictive_modelling/{model_name.lower().replace(\" \", \"_\")}_errors.png', dpi=150, bbox_inches='tight')\n",
     "    plt.show()\n"
    ]
   },

requirements.txt

@@ -6,5 +6,4 @@ scikit-learn
 jupyter
 notebook
 torch
-sqlite3
-boardlib
+boardlib