index.md

title	microsoftml API
description	microsoftml API
author	VanMSFT
ms.author	vanto
ms.date	07/15/2019
ms.service	sql
ms.subservice	machine-learning-services
ms.topic	reference
keywords	microsoftml API API
ms.devlang	python
monikerRange	>=sql-server-2017||>=sql-server-linux-ver15

microsoftml API

learners

Applies to: SQL Server 2017 RC2

training functions

• microsoftml.rx_fast_forest: Random Forest

• microsoftml.rx_fast_linear: Linear Model with Stochastic Dual Coordinate Ascent

  • loss functions

    • microsoftml.hinge_loss: Hinge loss function

    • microsoftml.log_loss: Log loss function

    • microsoftml.smoothed_hinge_loss: Smoothed hinge loss function

    • microsoftml.squared_loss: Squared loss function

• microsoftml.rx_fast_trees: Boosted Trees

• microsoftml.rx_logistic_regression: Logistic Regression

• microsoftml.rx_neural_network: Neural Network

  • optimizers

    • microsoftml.adadelta_optimizer: Adaptive learning rate method

    • microsoftml.sgd_optimizer: Stochastic gradient descent

  • math

    • microsoftml.avx_math: Acceleration with AVX instructions

    • microsoftml.clr_math: Acceleration with .NET math

    • microsoftml.gpu_math: Acceleration with NVidia CUDA

    • microsoftml.mkl_math: Acceleration with Intel MKL

    • microsoftml.sse_math: Acceleration with SSE instructions

• microsoftml.rx_oneclass_svm: Anomaly Detection

transforms

categorical variable handling

• microsoftml.categorical: Converts a text column into categories

• microsoftml.categorical_hash: Hashes and converts a text column into categories

schema manipulation

• microsoftml.concat: Concatenates multiple columns into a single vector

• microsoftml.drop_columns: Drops columns from a dataset

• microsoftml.select_columns: Retains columns of a dataset

variable selection

• microsoftml.count_select: Feature selection based on counts

• microsoftml.mutualinformation_select: Feature selection based on mutual information

text analytics

• microsoftml.featurize_text: Converts text columns into numerical features

  • N-grams extractors

    • microsoftml.n_gram: Converts text into features using n-grams

    • microsoftml.n_gram_hash: Converts text into features using hashed n-grams

  • Stopwords removers

    • microsoftml.custom: Removes custom stopwords

    • microsoftml.predefined: Removes predefined stopwords

• microsoftml.get_sentiment: Sentiment analysis

image analytics

• microsoftml.load_image: Loads an image

• microsoftml.resize_image: Resizes an Image

• microsoftml.extract_pixels: Extracts pixels form an image

• microsoftml.featurize_image: Converts an image into features

scorers

• microsoftml.rx_predict: Scores using a Microsoft machine learning model

featurizers

• microsoftml.rx_featurize: Data transformation for data sources