Scribd
Upload
5 views8 pages

Model Evaluation and Regression Techniques

The document discusses the evaluation and tuning of machine learning models, particularly focusing on accuracy, precision, and recall metrics. It highlights the importance of cross-validation and model selection to ensure effective performance. Additionally, it touches on regression analysis and the significance of estimating parameters for predicting trends in data.

Uploaded by

hejjdh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
5 views8 pages

Model Evaluation and Regression Techniques

The document discusses the evaluation and tuning of machine learning models, particularly focusing on accuracy, precision, and recall metrics. It highlights the importance of cross-validation and model selection to ensure effective performance. Additionally, it touches on regression analysis and the significance of estimating parameters for predicting trends in data.

Uploaded by

hejjdh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

TraindataCx YI

Train medll
Test data Cx Y
test
evaluate train model For certain
test data 只能 evaluate
不能 来 for tan nNN
not means
wins
all
aaracy
选 model win

Tuning
model compare medel
validn select
利 Cross
ti on
这项 枝 的
评估 INN
2NN 选那个 最好

U
Validation
data
Problems with accuracy
model useless
Wrong
99.97
even
accuracy
这个 model 法 判断
我们 想 观察 的 变重
also need to see

and recall
Precision
Precision
对 的 数 年个 variable

nunhnoliaf
畆
recall
在了
precision
预测 positive
做 分 crow
recall
真正 positive
做 分 c column

cul old a
Cross validation
steps
feld doesn't relate

kn [Link]
Cdtaiiǎgr foldnumber strata Y
create workflow
workflow c 1
ǎǎdiiiuǎǜij
[Link] [Link]
or tune
C gridc
re sample grid1
计算 acwacy 版
选 10个 同 k
Q metrics truth Y estimate prod_class
使 training data计算 accuracy1 vfold
和
cvcl li
tire sample计算有什么不同
只能 c 选 model
1 a

Regression ii
i
X nu regression
nn
Y nu

Model re ynu
model kun

Regression
对应
个x g
个

in
classification
在任意
阴彩 确定 y
住

-
linear regression
纯
atbxmodel estimate a 与b
lit
n 对 data
cxiyi
regression
a b
Estimate 可以
method
beast square 个点 tǎǎaǎi
住 是 每
画很多 条 线
起来 difference越
垂线 加
越好
[Link] bxi12
SSE 品
squared
Estmol
is
m
ㄑ

1 䈚髊
0
m RMSETSE in least
点
er
square
estimatea
nps samples
[Link]
for square b
estimate
Advantages

easy
internet duo
efficient
Computation
less data trends
need
minor
ancaunot
capture

linear Re
llubtiple
multiplex

at bxitcxztdxs
Why regression linear
不合
bend
当 trend
造

Goodness of fitSE MSE


RM

power test
prediction
dansiliatiaàuuraey
data
testing
RMSPE

You might also like