Skip to main content

​
SimpleLogisticModel

A simple logistic regression model implementation using gradient descent.

​
Constructor

SimpleLogisticModel(lr: float = 0.01, epochs: int = 600)
​
lr
float
default:"0.01"
Learning rate for gradient descent optimization
​
epochs
int
default:"600"
Number of training iterations

​
Attributes

​
weights
np.ndarray | None
Model weights including bias term. First element is bias, remaining are feature weights.
​
feature_columns
list[str] | None
Names of features used during training

​
Methods

​
fit

Train the logistic regression model on provided data. 
fit(X: pd.DataFrame, y: pd.Series) -> SimpleLogisticModel
​
X
pd.DataFrame
required
Feature matrix with samples as rows and features as columns
​
y
pd.Series
required
Binary target labels (0 or 1)
​
return
SimpleLogisticModel
Returns self for method chaining

​
predict_proba

Predict class probabilities for samples. 
predict_proba(X: pd.DataFrame) -> np.ndarray
​
X
pd.DataFrame
required
Feature matrix for prediction. Must contain all feature columns from training.
​
return
np.ndarray
Array of shape (n_samples, 2) with probabilities for class 0 and class 1

​
predict

Predict binary class labels for samples. 
predict(X: pd.DataFrame) -> np.ndarray
​
X
pd.DataFrame
required
Feature matrix for prediction. Must contain all feature columns from training.
​
return
np.ndarray
Binary predictions (0 or 1) using 0.5 probability threshold

​
ModelArtifacts

Dataclass containing trained models and test data.
​
risk_model
SimpleLogisticModel
Trained model for risk prediction
​
outcome_model
SimpleLogisticModel
Trained model for outcome prediction
​
X_test
pd.DataFrame
Test features
​
y_risk_test
pd.Series
Test labels for risk target
​
y_outcome_test
pd.Series
Test labels for outcome target

​
train_predictive_models

Train risk and outcome prediction models with automatic preprocessing and train/test split. 
train_predictive_models(
    df: pd.DataFrame,
    feature_cols: list[str],
    risk_target: str,
    outcome_target: str
) -> ModelArtifacts
​
df
pd.DataFrame
required
Input dataset containing features, hospital, gender, and target columns
​
feature_cols
list[str]
required
List of numerical feature column names to use for modeling
​
risk_target
str
required
Column name for risk categorization (checks for “appendicitis” or “pregnancy”)
​
outcome_target
str
required
Column name for outcome target (binarized as ‘t’ vs others)
​
return
ModelArtifacts
Contains trained risk and outcome models plus test data for evaluation

​
evaluate_predictive_models

Evaluate trained models on test data with multiple metrics. 
evaluate_predictive_models(artifacts: ModelArtifacts) -> dict[str, float]
​
artifacts
ModelArtifacts
required
Model artifacts from train_predictive_models containing models and test data
​
return
dict[str, float]
Dictionary containing:
  • risk_accuracy: Accuracy for risk model
  • risk_f1: F1 score for risk model
  • risk_auc: AUC-ROC for risk model
  • outcome_accuracy: Accuracy for outcome model
  • outcome_f1: F1 score for outcome model
  • outcome_auc: AUC-ROC for outcome model
  • sample_count: Number of test samples

Build docs developers (and LLMs) love

Get started for freeTalk to us