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