Supervised Numeric-Value Prediction Task

A supervised numeric-value prediction task is a data-driven point estimation task that is a supervised estimation task.

• Context:
  • Input: a Numerically Labeled Training Dataset.
  • output: an Predicted Numeric Value.
    • optional output: a Trained Estimation Function.
  • Performance Metric: Numeric-Value Prediction Performance Metric, such as RMSE.
  • It can be solved by a Supervised Numeric Prediction System (that implements a supervised point estimation algorithm).
  • It can range from being a Supervised Model-based Numeric Prediction Task (such as linear regression) to being a Supervised Instance-based Point Estimation Task (such as k-nearest regression).
  • It can range from being a Fully-Supervised Regression Task to being a Semi-Supervised Regression Task.
  • It can range from being a Univariate Regression Task to being a Multivariate Regression Task (with more than one dependent variable).
• Example(s):
  • predict the orbit of stellar objects based on past history (like Gauss did on ~1805).
  • predict Boston housings sales price based on a (Harrison & Rubinfeld, 1978) dataset, such as sklearn's Boston dataset.
  • a (Fanaee-T & Gama, 2014) Prediction Task.
  • …
• Counter-Example(s):
  • a Heuristic Point Estimation Task.
  • a Supervised Classification Task.
  • a Supervised Range Estimation Task.
  • a Supervised Ranking Task.
• See: Function Fitting; Residual; Kernel Function; Point Estimator; Gaussian Processes; Linear Regression; Support Vector Machines; Regression Analysis.

---

---

References

• http://en.wikipedia.org/wiki/Regression_analysis

2011

• (Quadrianto & Buntine, 2011) ⇒ Novi Quadrianto; Wray I. Buntine. (2011). “Regression.” In: (Sammut & Webb, 2011) p.838

1997

• (Mitchell, 1997) ⇒ Tom M. Mitchell. (1997). “Machine Learning." McGraw-Hill.
  • QUOTE: Much of the literature on nearest-neighbor methods and weighted local regression uses a terminology that has arisen from the field of statistical pattern recognition. ...
    • Regression means approximating a real-valued target function.
    • Residual is the error f^{^}(x) - [math]\displaystyle{ f }[/math](x) in approximating the target function.
    • Kernel function is the function of distance that is used to determine the wight of each training example. In other words, the kernel function is the function [math]\displaystyle{ K }[/math] such that w_i = K(d(x_i, x_q)).

---