Model-based Supervised Numeric-Value Prediction Task

A Model-based Supervised Numeric-Value Prediction Task is a supervised numeric-value prediction task that is a model-based supervised learning task.

• AKA: Continuous Function Regression, Supervised Regression Task.
• Context:
  • input:
    • a Regression Task Input.
    • a Real-Value Function Family.
  • output: Fitted Real-Value Function.
  • It can range from being a Model-based Parametric Estimation Task to being a Model-based Non-Parametric Point Estimation Task.
  • It can range from being a Fully-Supervised Model-based Regression Task to being a Semi-Supervised Model-based Regression Task.
  • It can range from being a Linear Regression Task to being a Non-Linear Regression Task.
  • It can be solved by a Model-based Supervised Numeric-Value Prediction System (that implements a model-based supervised numerical-value prediction algorithm).
• Example(s):
  • a Statistical Regression Task.
  • a model-based Supervised House Price Prediction Task, such as: [1]
  • a Least-Squares Point Estimation Task.
  • a Regression Tree Training Task.
  • a Maximum Entropy-based Regression Task.
  • …
• Counter-Example(s):
  • Supervised Classification Function Learning Task, such as a logistic function fitting task.
  • Supervised Model-based Classification Task.
  • Instance-based Supervised Point Estimation Task.
  • a Nonparametric Regression Task.
• See: Regression Model, Supervised Machine Learning Task, Supervised Machine Learning Task, Dynamic Programming.

---

---

References

2011

• http://en.wikipedia.org/wiki/Linear_least_squares_%28mathematics%29#Motivational_example
  • As a result of an experiment, four [math](x, y)[/math] data points were obtained, [math](1, 6),[/math] [math](2, 5),[/math] [math](3, 7),[/math] and [math](4, 10)[/math] (shown in red in the picture on the right). It is desired to find a line [math]y=\beta_1+\beta_2 x[/math] that fits "best" these four points. In other words, we would like to find the numbers [math]\beta_1[/math] and [math]\beta_2[/math] that approximately solve the overdetermined linear system [math]\begin{alignat}{3} \beta_1 + 1\beta_2 &&\; = \;&& 6 & \\ \beta_1 + 2\beta_2 &&\; = \;&& 5 & \\ \beta_1 + 3\beta_2 &&\; = \;&& 7 & \\ \beta_1 + 4\beta_2 &&\; = \;&& 10 & \\ \end{alignat}[/math] of four equations in two unknowns in some "best" sense.

2006

• (Tibshirani, 1996) ⇒ Robert Tibshirani. (1996). “Regression Shrinkage and Selection via the Lasso.” In: Journal of the Royal Statistical Society, Series B, 58(1).
  • QUOTE: Consider the usual regression situation: we have data [math](\mathbf{x}^i, y^i), i=1,2,...,N \ ,[/math] where [math]\mathbf{x}^i=(x_{i1},..., x_{ip})^T[/math] and [math]y_i[/math] are the regressors and response for the ith observation. The ordinary least squares (OLS) estimates are obtained by minimizing the residual squared error.

---