PATTERN RECOGNITION
AND MACHINE LEARNING
CHAPTER 3: LINEAR MODELS FOR REGRESSION
Linear Basis Function Models (1)
Example: Polynomial Curve Fitting
确定映射关系
Sum-of-Squares Error Function
确定目标函数
确定 w 使目标函数最小化
0th Order Polynomial
1st Order Polynomial
3rd Order Polynomial
9th Order Polynomial
避免过拟合： 1 、增加训练集 2 、增加惩罚项
Linear Basis Function Models (2)
Generally
where Áj(x) are known as basis functions.
Typically, Á0(x) = 1, so that w0 acts as a bias.
In the simplest case, we use linear basis
functions : Ád(x) = xd.
Linear Basis Function Models (3)
Polynomial basis functions:
These are global; a small
change in x affect all basis
functions.
Linear Basis Function Models (4)
Gaussian basis functions:
These are local; a small change
in x only affect nearby basis
functions. ¹j and s control
location and scale (width).
Linear Basis Function Models (5)
Sigmoidal basis functions:
where
Also these are local; a small
change in x only affect nearby
basis functions. ¹j and s control
location and scale (slope).
Maximum Likelihood and Least Squares (1)
Assume observations from a deterministic function
with added Gaussian noise:
方差
where
which is the same as saying,
Given observed inputs, , and targets,
, we obtain the likelihood function
最大化似然函数
Maximum Likelihood and Least Squares (2)
在高斯噪声下，才能写成该形
Taking the logarithm, we get 式，此时最小均方误差和最大
似然估计等价
where
is the sum-of-squares error.
Maximum Likelihood and Least Squares (3)
Computing the gradient and setting it to zero yields
Solving for w, we get The Moore-Penrose
pseudo-inverse, .
where
多个（ x,y ）对应
Geometry of Least Squares
Consider Φ 相当于空间中的基，称其为基函数
N-dimensional
M-dimensional
S is spanned by .
wML minimizes the distance
between t and its orthogonal
projection on S, i.e. y.
Y 的每一个值为 M 个向量的线型组合
Sequential Learning
Data items considered one at a time (a.k.a.
online learning); use stochastic (sequential)
gradient descent:
This is known as the least-mean-squares (LMS)
algorithm. Issue: how to choose ´?
1st Order Polynomial
3rd Order Polynomial
9th Order Polynomial
Over-fitting
Root-Mean-Square (RMS) Error:
Polynomial Coefficients
Regularized Least Squares (1)
Consider the error function:
添加惩罚项
Data term + Regularization term
With the sum-of-squares error function and a
quadratic regularizer, we get
¸ is called the
which is minimized by regularization
coefficient.
Regularized Least Squares (1)
Is it true that or
Regularized Least Squares (2)
With a more general regularizer, we have
Lasso Quadratic
Regularized Least Squares (3)
Lasso tends to generate sparser solutions than a
quadratic
regularizer.
Understanding Lasso Regularizer
• Lasso regularizer plays the role of
thresholding
Multiple Outputs (1)
Analogously to the single output case we have:
Given observed inputs, , and targets,
, we obtain the log likelihood function
Multiple Outputs (2)
Maximizing with respect to W, we obtain
If we consider a single target variable, tk, we see that
where , which is identical with the
single output case.
The Bias-Variance Decomposition (1)
Recall the expected squared loss,
where
The second term of E[L] corresponds to the noise
inherent in the random variable t.
What about the first term?
The Bias-Variance Decomposition (2)
Suppose we were given multiple data sets, each of
size N. Any particular data set, D, will give a
particular function y(x;D). We then have
The Bias-Variance Decomposition (3)
Taking the expectation over D yields
偏差描述与真实值之间的差距
方差描述与观测值均值之间的差距
The Bias-Variance Decomposition (4)
Thus we can write
where
Bias-Variance Tradeoff
The Bias-Variance Decomposition (5)
Example: 25 data sets from the sinusoidal, varying
the degree of regularization, ¸.
The Bias-Variance Decomposition (6)
Example: 25 data sets from the sinusoidal, varying
the degree of regularization, ¸.
The Bias-Variance Decomposition (7)
Example: 25 data sets from the sinusoidal, varying
the degree of regularization, ¸.
The Bias-Variance Trade-off
From these plots, we note
that an over-regularized
model (large ¸) will have a
high bias, while an under-
regularized model (small ¸)
will have a high variance.
Bayesian Linear Regression (1)
Define a conjugate prior over w
Combining this with the likelihood function and using
results for marginal and conditional Gaussian
distributions, gives the posterior
where
Bayesian Linear Regression (2)
A common choice for the prior is
for which
Next we consider an example …
Bayesian Linear Regression (3)
0 data points observed
Prior Data Space
Bayesian Linear Regression (4)
1 data point observed
Likelihood Posterior Data Space
Bayesian Linear Regression (5)
2 data points observed
Likelihood Posterior Data Space
Bayesian Linear Regression (6)
20 data points observed
Likelihood Posterior Data Space
Predictive Distribution (1)
Predict t for new values of x by integrating
over w:
where
Predictive Distribution (1)
How to compute the predictive distribution ?
Predictive Distribution (2)
Example: Sinusoidal data, 9 Gaussian basis functions,
1 data point
Predictive Distribution (3)
Example: Sinusoidal data, 9 Gaussian basis functions,
2 data points
Predictive Distribution (4)
Example: Sinusoidal data, 9 Gaussian basis functions,
4 data points
Predictive Distribution (5)
Example: Sinusoidal data, 9 Gaussian basis functions,
25 data points
Salesman’s Problem
Given
• a consumer a described by a vector x
• a product b to sell with base cost c
• estimated price distribution of b in the mind
of a is Pr(t|x, w) = N(xTw, 2)
What is price should we offer to a ?
Salesman’s Problem
Bayesian Model Comparison (1)
How do we choose the ‘right’ model?
Bayesian Model Comparison (1)
How do we choose the ‘right’ model?
Assume we want to compare models Mi, i=1, …,L,
using data D; this requires computing
Posterior Prior Model evidence or
marginal likelihood
Bayesian Model Comparison (1)
How do we choose the ‘right’ model?
Posterior Prior Model evidence or
marginal likelihood
Bayesian Model Comparison (3)
For a model with parameters w, we get the
model evidence by marginalizing over w
Bayesian Model Comparison (4)
For a given model with a
single parameter, w, con-
sider the approximation
where the posterior is
assumed to be sharply
peaked.
Bayesian Model Comparison (4)
For a given model with a
single parameter, w, con-
sider the approximation
where the posterior is
assumed to be sharply
peaked.
Bayesian Model Comparison (5)
Taking logarithms, we obtain
Negative
With M parameters, all assumed to have the same
ratio , we get
Negative and linear in M.
Bayesian Model Comparison (5)
Data matching Model Complexity
Bayesian Model Comparison (6)
Matching data and model complexity
What You Should Know
• Least square and its maximum likelihood
interpretation
• Sequential learning for regression
• Regularization and its effect
• Bayesian regression
• Predictive distribution
• Bias variance tradeoff
• Bayesian model selection
