Franklin Moormann issues

Results 113 issues of


                                            Franklin Moormann

[Phase 2 Parity] Implement AdamW and Learning Rate Schedulers

Problem: Missing AdamW and all LR schedulers. Missing: AdamW (HIGH), StepLR (HIGH), CosineAnnealingLR (HIGH), ReduceLROnPlateau (HIGH), ExponentialLR (HIGH), CyclicLR (MEDIUM). Existing: Adam, SGD, RMSprop, Adagrad in src/Optimizers/. Architecture: src/Optimizers/Schedulers/. Goal:...

[Phase 2 Parity] Implement Critical Preprocessing Transformers

Problem: Missing fundamental sklearn preprocessing. Missing: OneHotEncoder (CRITICAL), LabelEncoder (CRITICAL), PolynomialFeatures (HIGH), SimpleImputer (HIGH), KNNImputer (MEDIUM), PowerTransformer (MEDIUM). Existing: StandardScaler, MinMaxScaler, RobustScaler in src/Normalizers/. Architecture: src/DataProcessor/Transformers/. Goal: Fit/Transform pattern, pipeline...

[Phase 2 Parity] Implement Ridge and Lasso Regression

Problem: No L1/L2 regularized linear regression. Missing: Ridge (HIGH), Lasso (HIGH), ElasticNet (HIGH), RidgeCV, LassoCV, ElasticNetCV. Use Cases: Feature selection, multicollinearity handling, high-dimensional data. Goal: Efficient coordinate descent solver, alpha...

[Phase 2 Parity] Implement Naive Bayes Classifiers

Problem: No Naive Bayes implementations. Missing: GaussianNB (CRITICAL), MultinomialNB (CRITICAL), BernoulliNB (CRITICAL), ComplementNB (MEDIUM). Use Cases: Text classification, spam filtering, real-time prediction. Architecture: src/Classification with NaiveBayesBase. Goal: Partial fit support,...

[Phase 2 Parity] Implement Support Vector Machines Classifier

Problem: AiDotNet has SVR but lacks SVC. Missing: SVC (CRITICAL), LinearSVC (HIGH), NuSVC (MEDIUM), OneClassSVM (MEDIUM). Architecture: src/Classification/ with IClassifier interface. Goal: Multi-class support with probabilistic outputs, parity with sklearn.svm.

[Phase 2 Parity] Implement Modern Dimensionality Reduction Algorithms

Problem: AiDotNet has decomposition methods but lacks high-level dimensionality reduction. Missing: t-SNE (CRITICAL), UMAP (HIGH), PCA wrapper (HIGH), LDA (HIGH), ICA (MEDIUM). Existing: SVD and Eigen decomposition. Goal: Build high-level...

[Phase 2 Parity] Implement Comprehensive Clustering Algorithms

## Problem AiDotNet lacks standard clustering algorithms that are fundamental in scikit-learn. ## Missing Implementations **CRITICAL Priority:** - KMeans clustering - DBSCAN (Density-Based Spatial Clustering) **HIGH Priority:** - Hierarchical Clustering...

Franklin Moormann