import { RegressorBase } from '../base';
import { Params } from '../base/estimator';
export interface GradientBoostingRegressorProps {
    nEstimators?: number;
    n_estimators?: number;
    learningRate?: number;
    learning_rate?: number;
    maxDepth?: number;
    max_depth?: number;
    minSamplesSplit?: number;
    min_samples_split?: number;
    subsample?: number;
    maxFeatures?: number | 'sqrt' | 'log2';
    max_features?: number | 'sqrt' | 'log2';
    randomState?: number;
    random_state?: number;
}
/**
 * Gradient boosting with squared-error loss, following sklearn's
 * GradientBoostingRegressor: F_0 = mean(y), then each round fits a
 * regression tree to the residuals y - F and updates
 * F += learning_rate * tree(x). subsample < 1 draws rows without
 * replacement per round (stochastic gradient boosting); trees are still
 * used to update F over all samples.
 */
export declare class GradientBoostingRegressor extends RegressorBase {
    private nEstimators;
    private learningRate;
    private maxDepth;
    private minSamplesSplit;
    private subsample;
    private maxFeatures?;
    private randomState?;
    private estimators;
    private initPrediction;
    private fitted;
    constructor(props?: GradientBoostingRegressorProps);
    getParams(): Params;
    fit(trainX: number[][], trainY: number[]): void;
    predict(testX: number[][]): number[];
    get featureImportances(): number[];
}