Machine learning, a branch of artificial intelligence that relate any system capable of acquiring and integrating the knowledge automatically. It is possible due to the capability of the systems to learn from training, analytical observation, experience, and other means, which leads to a system that can continuously self-improve and hence can enhance effectiveness efficiency.
Why Machine Learning?
- Human produce machines that do not work as well as desired in the environments in which they are used.
- Time is not static. Environments change over time.
- Some tasks cannot be defined well, except by examples.
- Relationships and correlations can be hidden within large amounts of data. Machine Learning/Data Mining may be able to find these relationships.
- The Amount of knowledge available about certain tasks might be large for encoding by humans.
- New scoop about tasks is regularly being discovered by humans. So It may be difficult to continuously re-design systems by hand.
Various Algorithm types
Machine learning algorithms can be divided into following types
- Supervised Learning: It generates a function that maps inputs to desired output. Supervised learning in machine learning software consists of a carefully calculated set of data, to show the software program the correct patterns to follow. The program must then manipulate the input data in order to fit the patterns given by the trainer, and make adjustments as directed by the trainer. During
- Unsupervised Learning: Unsupervised learning is more difficult: the aim is to have the computer learn how to do something that we don’t tell it how to do! .In this, the computer program is given the set of data and patterns but must make adjustments itself without direction. The program learns itself without human intervention by trial and error.
- Semi-Supervised Learning: As the name suggest, this approach combines both labeled and unlabeled instructions to generate an appropriate function or classifier.
- Reinforcement Leaning: This approach learns how to act given an observation of the world. Since every action has some impact in the environment, and that environment gives feedback in the form of rewards that guides the self-learning algorithm.
- Transduction: It tries to predict new outputs based on training inputs, training outputs, and test inputs.
- Learning to learn: It learns its own inductive bias based on previous experience.
Various Examples of Successful Applications of Machine Learning
- Learning to recognize spoken words (Lee, 1989; Waibel, 1989).
- Learning to drive an autonomous vehicle (Pomerleau, 1989).
- Learning to classify new astronomical structures (Fayyad et al., 1995).
- Learning to play world-class backgammon (Tesauro 1992, 1995).
Also Read: Digital Jewellery
Machine learning application domains:
Machine learning and other statistical tools can be used to predict stock market fluctuations.
Machine learning tools have been used to predict the three-dimensional structure of proteins derived from gene sequences.
An early application of neural networks was to identify certain vehicle types (such as tanks) from visual information.
Predictive toxicology (described below) uses machine learning agents to decide which drugs may turn out to be toxic to humans.
- Natural language
Agents are programmed to learn grammars in order to improve natural language comprehension.
Diagnosing patients based on their symptoms is an important application of machine learning tools.
Recent applications to computer music include the use of Markov models to classify music into styles