Microsoft Push Azure Machine learning tool: Algorithm Cheat Sheet

Microsoft Push Azure Machine learning tool: Algorithm Cheat Sheet

[Date: 2015-05-15]

Source: CSDN Linux

[Font: Big Small]

Azure Machine Learning Studio has a number of machines learning algorithms that you can now use to build predictive analytics solutions. These algorithms can be used for general machine learning: regression analysis, classification, clustering, and anomaly detection , each of which solves different types of machine learning problems.

The question now is, is there any tool or something that can help find out how to choose a suitable machine learning algorithm and according to the specific scenario?

Click here to see a larger image.

Click here to download Cheat Sheet:microsoft Azure machine learning algorithm Cheat Sheet

Microsoft Azure's machine learning algorithm Cheat Sheet is designed to help you filter available machine learning algorithms and choose the right one for predictive analytics solutions. Cheat Sheet will ask you these two questions: the nature of the data, the problem you want to work with, and then suggest an algorithm that you can try out.

Azure machine learning Studio gives you a flexible experience: try an algorithm, and if you're not happy with the results, try another. (Azure machine learning is free and does not require licensing, click here to try.) Here is an example from Azure machine learning Gallery, which attempts to use several different algorithms on the same data and then compare the results: Compare multi-class Classifiers:letter Recognition.

• For a more in-depth discussion of the different types of machine learning algorithms and how they are used, see: How to choose a algorithm in Azure machine learning.
• The list of all machine learning algorithms is available in machines learning Studio, see: Initialize Model.

About Azure Machine learning Studio

Azure Machine Learning Studio offers a number of different advanced machines learning algorithms to help you generate analytic models. First, determine the general type of machine learning task that you want to perform, because the algorithms that are grouped within each category are appropriate for a specific forecast task.

After you select an algorithm and configure its parameters, you can use one of the training modules to run the data through the selected algorithm, or you can use the scan parameters to iterate through all possible parameters and determine the optimal configuration of tasks and data.

Types of learning Algorithms

Azure Machine Learning Studio provides the following kinds of machines learning algorithms, grouped by typical machine learning scenarios.

• Anomaly detection

Anomaly detection includes many important tasks in machine learning, and anomaly detection technology is suitable for a variety of industries:

By definition, exceptions are rare events, so it may be difficult to collect representative data samples for modeling. The algorithms included in this section have been specifically designed to address the core building and training model issues of anomaly detection.

This category includes the following modules: single-class support vector machines, PCA-based anomaly detection.

• Classification

The classification algorithm is used to predict the class or category of a single data instance. For example, an e-mail filter uses a two-tuple classification to determine whether an e-mail message is spam. There are two types of classification tasks. One is the $ two classification that is intended to predict one of the two outcomes, and the other is a multi-class classification that predicts one of several outcomes. The output of the classification algorithm is a classifier that can be used to predict the label of a new (unlabeled) instance.

Category Modules References.machine learning.initialize model.classification includes the following modules: Multi-class decision forest, multi-class decision forest, multi-class logistic regression, multi-class neural network, a pair of Toto, Two-class average Perceptron, two-class Bayesian point machine, two-class elevation decision tree, two-type decision forest, two-class decision forest, two-class logistic regression, two-class neural network, two-class support vector machine, two-class local deep support vector machine.

• Clustering

Clustering algorithms can be based on a set of features to learn how to group items together. For example, clustering is typically used in text analysis to group text fragments that contain common words. You can use clustering to group unlabeled data by finding the closest data point and then determining the centroid or center point of each combination. After the algorithm is trained, it can be used to predict which cluster the data instance belongs to.

Category Modules References.machine learning.initialize model.clustering including modules: K-means clustering

• Regression

A regression algorithm is an algorithm that learns the actual function words that predict a single data instance. For example, a house price predictor can use a regression algorithm to predict current house prices. The regression algorithm determines the distribution of each feature of the data to perform the regression function. After the algorithm trains a function that predicts the marking of data, it can be used to predict the label of a new (unlabeled) instance.

Category Modules References.machine learning.initialize model.regression

English Original: Microsoft Azure machine learning algorithm Cheat Sheet

This article permanently updates the link address : http://www.linuxidc.com/Linux/2015-05/117567.htm

Microsoft Push Azure Machine learning tool: Algorithm Cheat Sheet