Unsupervised Learning Task

Context:
- Input: Learning Data Records, a Dataset X (without target out put y).
- output: Model that Predicts a Test Case's Cluster.
- It can be solved by an Unsupervised Learning System (that implements an unsupervised learning algorithm).
Example(s):
- a Data-Driven Clustering Task.
- a Database Segmentation Task.
- an Unsupervised Classification Task.
- a Generative Adversarial Networks Task.
- …
Counter-Example(s):
- a Supervised Learning Task, such as supervised class prediction or supervised numeric prediction.
- a Reinforcement Learning Task.
See: Probability Density Estimation, Exploratory Data Analysis, Perception Statistical Learning, Deep Learning.

References

(Wikipedia, 2019) ⇒ https://en.wikipedia.org/wiki/unsupervised_learning Retrieved:2019-12-4.
- Unsupervised learning is a type of self-organized Hebbian learning that helps find previously unknown patterns in data set without pre-existing labels. It is also known as self-organization and allows modeling probability densities of given inputs. It is one of the main three categories of machine learning, along with supervised and reinforcement learning. Semi-supervised learning has also been described, and is a hybridization of supervised and unsupervised techniques. ...

(Sammut & Webb,2017) ⇒ Claude Sammut, Geoffrey I. Webb. (2017)."Unsupervised Learning" In: Encyclopedia of Machine Learning and Data Mining pp 1304-1304,
- QUOTE: Unsupervised learning refers to any machine learning process that seeks to learn structure in the absence of either an identified output (cf. supervised learning) or feedback (cf. reinforcement learning). Three typical examples of unsupervised learning are clustering, association rules, and self-organizing maps.

(Triplet & Foucher, 2017) ⇒ Thomas Triplet, and Samuel Foucher (2017). "Clustering of Geospatial Big Data in a Distributed Environment". In: Encyclopedia of GIS pp 236-246
- QUOTE: Clustering, sometimes called unsupervised learning/classification or exploratory data analysis, is one of the most fundamental steps in understanding a dataset, aiming to discover the unknown nature of data through the separation of a finite dataset, with little or no ground truth, into a finite and discrete set of “natural,” hidden data structures. Given a set of n points in a two-dimensional space, the purpose of clustering is to group them into a number of sets based on similarity measures and distance vectors. Clustering is also useful for compression purpose in large databases (Daschiel and Datcu 2005). The term Unsupervised Learning is sometimes used in some fields (i.e., in Machine Learning and Data Mining). Clustering will usually aim at creating homogeneous groups that are maximally separable. It is a fundamental tool in Knowledge Discovery and Data (KDD) mining when looking for meaningful patterns (Alam et al. 2014). Geographical Knowledge Discovery (GKD) is seen as an extension of KDD to the case of spatial data (Miller 2010).

(Sebag, 2011) ⇒ Michele Sebag, M. (2011). "Nonstandard Criteria in Evolutionary Learning". In Encyclopedia of Machine Learning (Sammut & Webb, 2011, pp. 722-731). Springer US.
- QUOTE: Machine learning (ML), primarily concerned with extracting models or hypotheses from data, comes into three main flavors: supervised learning also known as classification or regression (Bishop 2006; Duda et al. 2001; Han and Kamber 2000), unsupervised learning also known as clustering (Ben-David et al. 2005), and reinforcement learning(Sutton and Barto 1998).

(Binder et al.,2009) ⇒ Marc D. Binder, Nobutaka Hirokawa, and Uwe Windhorst.(2009)."Unsupervised Learning". In: Encyclopedia of Neuroscience pp 4154-4154
- QUOTE: Unsupervised learning is a form of learning in computational models such as connectionist (artificial neural network) models. In contrast to supervised learning, unsupervised learning algorithms work without providing explicit feedback on the error of the net with respect to its input (i.e., no teaching signal). Learning develops by using internal or statistical structure of data set, so that the responses (output) will be fully characterized statistical properties of inputs. Often, the aim of unsupervised learning algorithms is to cluster the input according to similarity. While this is biologically more plausible than providing an external teaching signal, problematic issues in this context are how many clusters to form, and when to stop training. Often, weights are adjusted until some internal constraint is fulfilled. It has been proposed that unsupervised learning occurs in cortex-based learning.

(Redei, 2008) ⇒ George P. Rédei. (2008). "Unsupervised Learning". In: Encyclopedia of Genetics, Genomics, Proteomics and Informatics pp 2067-2067
- QUOTE: Identifies new, so far undetected, shared pattern(s) of sequences in macromolecules and determines the positive and negative representatives of the pattern(s). The information permits correlations between structure and function in languages as well as in proteins without direct human intervention in the details

(Valpola, 2000) ⇒ Harri Valpola. (2000). “Bayesian Ensemble Learning for Nonlinear Factor Analysis." PhD Dissertation, Helsinki University of Technology.
- QUOTE: unsupervise learning: The goal in unsupervised learning is to find an internal representation of the statistical structure of the observations. See supervised learning.

(Kohavi & Provost, 1998) ⇒ Ron Kohavi, and Foster Provost. (1998). “Glossary of Terms.” In: Machine Leanring 30(2-3).
- QUOTE: Unsupervised learning: Learning techniques that group instances without a pre-specified dependent attribute. Clustering algorithms are usually unsupervised.