Machinelearning: First, what is machine learning

Brief introduction

Before I introduce machine learning, I would like to start by listing some examples of machine learning:

• junk e-mail detection: Identifies what is spam and what is not, based on the messages in the mailbox. Such a model can help categorize spam and non-spam messages by programs. This example, we should not be unfamiliar.
• Credit card fraud detection: Identifies which transactions are the user's operation and which are not, based on the user's credit card transactions within one months. Such a decision model can help the program to return those fraudulent transactions.
• Digital Recognition: Identifies the number represented by each hand-written character according to the handwritten zip code on the envelope. Such a model can help the program to read and understand the handwritten zip code, and classify letters according to geographical location.
• speech recognition: From a user's discourse, determine the specific requirements of the user. Such a model can help the program to be able and try to automatically populate the user's needs. iphone with the Siri system has this feature.
• Face Recognition: Identify photos that contain a person, based on the many digital photos in the album. Such a decision model can help the program manage photos based on the face of the person. Some cameras or software, such as iphoto, have this feature.
  
• Product Recommendation: Identify which of these are the products that the user is genuinely interested in and willing to buy, based on a user's shopping history and lengthy list of favorites. Such a decision model can help the program to provide advice to customers and encourage product consumption. Sign in to Facebook or googleplus and they'll recommend users who might be associated with it.
• Medical Analysis: predict what disease the patient may be suffering from, based on the patient's symptoms and an anonymous patient data database. Such a decision-making model can provide support for professional medical professionals.
• Stock Exchange: Determine whether the stock is open, open or Jiancang according to the current and previous price fluctuations of a stock. Such a decision model can help the program provide support for financial analysis.
• Customer Segmentation: Identifies which users will be turned into the payment user for the product, and which will not, based on the behavior patterns of the user during the probation period and the behavior of all users in the past. Such a decision model can help the program with user intervention to persuade users to make early payments or better participate in product trials.
• Shape identification: Determines the shape the user wants to portray, based on the user's hand-drawn and a known shape library on the touch screen. Such a decision model can help the program display the ideal version of the shape to draw a clear image. iphone App Instaviz can do that. [1]

machine Learning (machines learning, ML) is a multidisciplinary interdisciplinary subject involving probability theory, statistics, approximation theory, convex analysis, algorithmic complexity theory and many other disciplines. Specialized in computer simulation or realization of human learning behavior, in order to acquire new knowledge or skills, reorganize the existing knowledge structure to continuously improve their performance. It is the core of artificial intelligence, mainly using induction, synthesis rather than deduction [2]

There are many definitions of machine learning, a frequently quoted definition of English: A computer program was said to learn from experience E with respect to some class of tasks T and Performance measure P, if its performance at the tasks in T, as measured by P, improves with experience e.[3]

How is this definition understood? Imagine a mailbox filter that accepts a message, how does the program determine if the message is spam? First of all, this email will first look at the emails you've flagged as spam, and based on this learning experience, the filter will filter out spam for you. In this task:

• Determine if the message is spam, this task T
• See if you've flagged a spam email before, this is experience e
• Correctly classify the number of spam and non-spam messages, this is the performance p

Machine Learning Classification

In the broadest sense, any method that involves machine learning can be thought of as using learning as long as it takes advantage of information from training samples. In practice and meaningful machine learning is so difficult that it is impossible to guess the best categorical decision. So most of the time is used to study the problem of learning. Implement a machine learning algorithm design: Given the form of a generic model or classifier, use training samples to learn or evaluate unknown parameters of the model. Learning here means using an algorithm to reduce the classification error of the training sample. A large class of algorithm based on gradient descent, can adjust the parameters of the classifier, so that it can reduce the error in the direction of moving forward, has become the mainstream learning algorithm in machine learning field. [4]
In a machine learning application, there are different ways to model a problem depending on the type of data. In the field of machine learning or artificial intelligence, people will first consider the algorithm's learning style. In the field of machine learning, there are several main ways of learning. It is a good idea to classify the algorithm according to the way of learning, which allows people to consider the best possible results by choosing the most suitable algorithm based on the input data when modeling and algorithm selection. [5]

Supervised learning

Under supervised learning, the input data is called "training data", each set of training data has a clear identification or results, such as the anti-spam system "spam" "non-spam", the handwritten numeral recognition of "1", "2", "3", "4" and so on. In the establishment of the predictive model, supervised learning establishes a learning process, compares the predicted results with the actual results of the "training data", and adjusts the predictive model continuously until the predicted results of the model reach an expected accuracy rate. Common application scenarios for supervised learning such as classification problems and regression problems. Common algorithms are: Logistic regression, perceptron, SVM support vector machine.

Non-supervised learning

In unsupervised learning, the data is not specifically identified, and the learning model is designed to infer some intrinsic structure of the data. Common application scenarios include learning about association rules and clustering. Common algorithms include: Apriori algorithm and K-means algorithm.

Semi-supervised learning

In this learning mode, the input data part is identified, the part is not identified, the learning model can be used for prediction, but the model first needs to learn the internal structure of the data in order to reasonably organize the data to make predictions. The application scenarios include classification and regression, and the algorithm includes some extensions to the commonly supervised learning algorithms, which first attempt to model the non-identified data, and then predict the identified data. On the inference algorithm (Graph inference) or Laplace support vector machine (Laplacian SVM).

Intensive Learning

In this learning mode, input data as feedback to the model, unlike the monitoring model, the input data is only as a check model of the wrong way, under the reinforcement learning, the input data directly feedback to the model, the model must be immediately adjusted. Common application scenarios include dynamic systems and robot control. Common algorithms include q-learning and time difference learning (temporal difference learning)

Machine Learning Framework

There are many frameworks, libraries and software in the Machine learning field. The more famous are: C + + OpenCV, shark, Java weka, Hadoop mahout, Spark Mllib,matlab LIBSVM, Caffe,python Scikit-learn, NLTK, Theano, Brain, Caffe and so on. Machine learning frameworks, libraries and software lists

Catalog (continuous update)

• Basic knowledge
• Probability distribution
• Regression
  
  
  • CART: Classification and regression tree algorithms
  • SVR Support vector regression
  
  
• Classification
  
  
  • Perception Machine
  • Logistic regression
    
    
    • Logistic regression
    • Logistic regression implementation
    • Logistic regression in the Scikit-learn
    
    
  • Classification decision number algorithm C4.5
  • Nearest neighbor Category
  • Naive Bayesian algorithm
  • CART: Classification and regression tree algorithms
  • Ada Boost iterative algorithm
  • Support Vector Machine
  • Graph model
  
  
• Clustering
  
  
  • K-mean-value clustering
  
  
• Time series
  
  
  • Time Series Full Tutorial (R)
  • Hmm hidden Markov model
  
  
• Dimension reduction
  
  
  • LDA Plain Understanding
  
  
• Other algorithms
  
  
  • Maximum expectation (EM) algorithm
  • Apriori algorithm
  
  
• Optimization algorithm
  
  
  • Gradient Descent method
  
  

#参考资料
[1] 10 Examples of machine learning
[2] Baidu Encyclopedia
[3] Machine learning
[4] Pattern classification
[5] Machine learning common classification

Machinelearning: First, what is machine learning