Machine Learning - العلم نور

جديد

{ وَقُلْ رَبِّ زِدْنِي عِلْمًا }

الجمعة، 28 يونيو 2019

Machine Learning


Machine Learning
      




   Machine Learning appeared to the world at the first time by Arthur Samuel in 1959. We can define machine learning simply as the technique by which machine can think, do tasks and make decisions. It's considered as one of the artificial intelligence branches, which can make decisions automatically by using algorithms. It's all about making the machine have a wide experience by having a learning data set for being able to deal with the new tasks accurately and make its own decisions.

Types of Machine Learning algorithms
         There are various types of machine learning algorithms, each type will be used according to the task which needs to be done. We can discuss these types as follows:
§  Supervised Learning_ It's used for building the model of data set mathematically according to the inputs and outputs of the system. The algorithm in this type receives a set of inputs and their corresponding outputs, this data is called training data which is used by the algorithm for comparing the correct outputs with the actual outputs for finding errors. A function is learned by the supervised learning algorithms for the prediction and determination of the output which corresponds to the new input. There are two parts of this type of algorithms i.e. classification and regression algorithms.
§  Semi-Supervised learning_ The only difference between supervised learning and semi-supervised learning is the usage of both labeled and unlabeled data for training. This type of learning is widely used with the too high labeling cost for the allowance of fully labeled training process.
§  Un-Supervised Learning_ It's the type of algorithms which is used for finding the structure of the set of data of which inputs only are known. The reaction of this type with the data set depends on identifying the commonalities in the data and checking for the absence or the presence of these commonalities in the new data.
§  Reinforcement Learning_ It's used specially for gaming, robotics and navigation. It's used for discovering the ways and actions which are expected to yield the greatest rewards. Agent, environment and actions are the main three components of reinforcement learning algorithm.








Machine Learning Applications
         Machine learning has a wide range of applications in almost all fields in everyday life. We will discuss the most important applications of machine learning in various fields.
ü Financial Services: Machine learning is used is financial field for the identification of data insights and also for the fraud prevention in banks and other businesses.
ü Health Care: Machine learning techniques are used widely in health care for the improvement of treatment and diagnosis by analyzing data and identifying new trends.
ü Oil and Gas Industry: Machine learning techniques are used widely in this field for the discovery of new energy sources, the predictions of sensors' failure and checking the ground minerals.
ü Transportation: Machine learning techniques are used in the transportation industry for the prediction of possible problems and the identification of new patterns and trends by analyzing the data in order to increase the efficiency of routes.

ليست هناك تعليقات:

إرسال تعليق