Home

Model Improvement

What you'll see here

Ten configurations across four levers, judged on validation recall first and F1 second.

The four levers
Depth / capacity

32 → 64-32 → 128-64-32 hidden units

Optimizer

SGD (lr 0.01) vs Adam (lr 0.001)

Regularisation

Dropout 0.2 and 0.3 between hidden layers

Imbalance handling

Class weights (9.01 : 0.53) and random oversampling

Every model uses the same split, same preprocessing, same 30 epochs and batch size 128, and a fixed seed of 42 — so differences come from the lever, not from run-to-run noise.

Model performance summary
ModelArchitectureOptimizerDropoutBalancingTrain recallVal accVal recallVal F1
M1 Baseline SGD32SGD-None59.12%97.6%58.56%73.03%
M2 SGD deeper64-32SGD-None74.66%98.58%75.23%85.42%
M3 Adam 2 layers64-32Adam-None90.43%99.25%90.54%93.06%
M4 Adam 3 layers128-64-32Adam-None92.23%99.08%87.84%91.33%
M5 Adam + dropout128-64-32Adam0.3None89.64%99.35%90.99%93.95%
M6 SGD + class weights64-32SGD-Class weights90.99%96.75%91.44%75.75%
M7 Adam + class weights128-64-32Adam-Class weights99.1%97.9%90.54%82.72%
M8 Adam + dropout + CW
Selected
128-64-32Adam0.3Class weights91.89%99.05%91.89%91.48%
M9 Adam + oversampling128-64-32Adam-Oversampling100%98.92%88.74%90.16%
M10 Adam + dropout + OS128-64-32Adam0.2Oversampling99.44%98.65%91.44%88.26%
Validation accuracy, recall and F1 across all ten configurations
Depth, optimizer and dropout (M1–M5)
  • Switching from SGD to Adam is the single biggest jump: recall rises from 58.6% to 90.5% at the same depth-2 architecture.
  • A third hidden layer alone (M4) overfits slightly — training recall 92.2% against validation 87.8%.
  • Adding dropout 0.3 to the deeper network (M5) closes that gap and delivers the best F1 of the unweighted models at 93.95%.
Class weights and oversampling (M6–M10)
  • Class weights lift recall in every case, but without regularisation they cost precision heavily — M6 drops to 64.7% precision, meaning one in three alarms is false.
  • M8 reaches the highest validation recall of the whole set at 91.89% while holding precision above 91%.
  • Oversampling reaches similar recall but memorises the duplicated minority rows — training recall of 99.4–100% against 88.7–91.4% on validation.
Imbalance handling compared — same 128-64-32 network
StrategyModelTrain recallVal recallVal precisionVal F1Read
NoneM589.64%90.99%97.12%93.95%Highest precision, misses most failures
Class weightsM891.89%91.89%91.07%91.48%Selected: best recall, precision held
OversamplingM1099.44%91.44%85.29%88.26%Same recall, clear memorisation gap
AI: Which lever mattered most

Click Generate to produce an evidence-based observation, insight, recommendation, and business benefit for this section.