def __init__(self, X, Y, kernel='LinearKernel', basis_vectors = None, regparams=None, **kwargs): if regparams is None: grid = [2**x for x in range(-15, 16)] else: grid = regparams learner = RLS(X, Y, grid[0], kernel, basis_vectors, **kwargs) crossvalidator = LPOCV(learner) self.cv_performances, self.cv_predictions, self.regparam = grid_search(crossvalidator, grid) self.predictor = learner.predictor
def __init__(self, X, Y, folds, kernel='LinearKernel', basis_vectors = None, regparams=None, measure=None, save_predictions = False, **kwargs): if regparams is None: grid = [2**x for x in range(-15, 16)] else: grid = regparams if measure is None: self.measure = sqerror else: self.measure = measure learner = RLS(X, Y, grid[0], kernel, basis_vectors, **kwargs) crossvalidator = NfoldCV(learner, measure, folds) self.cv_performances, self.cv_predictions, self.regparam = grid_search(crossvalidator, grid) self.predictor = learner.predictor
def __init__(self, X, Y, qids, kernel='LinearKernel', basis_vectors = None, regparams=None, measure=None, **kwargs): if regparams is None: grid = [2**x for x in range(-15, 15)] else: grid = regparams if measure is None: self.measure = cindex else: self.measure = measure learner = QueryRankRLS(X, Y, qids, grid[0], kernel, basis_vectors, **kwargs) crossvalidator = LQOCV(learner, measure) self.cv_performances, self.cv_predictions, self.regparam = grid_search(crossvalidator, grid) self.predictor = learner.predictor
def __init__(self, X, Y, kernel='LinearKernel', basis_vectors=None, regparams=None, measure=None, **kwargs): if regparams == None: grid = [2**x for x in range(-15, 16)] else: grid = regparams if measure == None: measure = sqerror learner = RLS(X, Y, grid[0], kernel, basis_vectors, **kwargs) crossvalidator = LOOCV(learner, measure) self.cv_performances, self.cv_predictions, self.regparam = grid_search( crossvalidator, grid) self.cv_predictions = np.array(self.cv_predictions) self.predictor = learner.predictor
def __init__(self, X, Y, qids, kernel='LinearKernel', basis_vectors=None, regparams=None, measure=None, **kwargs): if regparams is None: grid = [2**x for x in range(-15, 15)] else: grid = regparams if measure is None: self.measure = cindex else: self.measure = measure learner = QueryRankRLS(X, Y, qids, grid[0], kernel, basis_vectors, **kwargs) crossvalidator = LQOCV(learner, measure) self.cv_performances, self.cv_predictions, self.regparam = grid_search( crossvalidator, grid) self.predictor = learner.predictor