Collect typographical information from Toshiyuki Sakamoto's "Make and Understand! Introduction to Ensemble Learning Algorithms"

Toshiyuki Sakamoto "Making and Understanding! Introduction to Ensemble Learning Algorithms" (R & C, 2019)

--Publisher site: https://www.c-r.com/book/detail/1316

I looked at the publisher's site and found no typographical information anywhere, so I'll collect it.

page Wrong Positive
3 Is the purpose of this book I meanButThis is the purpose of this book
13 Normal. I use the decision tree as a base Normal**、**I use the decision tree as a base
36 「scikir-learn」 「scikit-learn」
40, 44 pred.argmax prob.argmax
43 Evaluation function in regression classification RegressionEvaluation function in
50~53 ncol nrow
57 Get worse results than Zero Rule With worse results than Zero RuleNanaRu
57 Prepare an array of the required size
62 If the number of dimensions and N Number of dimensionsToIf N
62 Because SGD can process data little by little SGD processes data little by littleBecause I can
68 Class variables Instance variables
81 The purpose of the decision tree algorithm Purpose of the decision tree algorithmis
85 Axis of data Of the dataNumber axis
88 Also create a "score" variable Also create a "score" variableI will
102 Difference left and right nodes Left and rightNodes
111 If you delete the branch in the node Delete the branch in the nodeif
117 def criticalscore(node, x, y, score_max) def criticalscore(node, score_max)
119 "Pruntest" which shows the ratio when the test data is separated Represents the percentage when the test data is separated**「splitratio」**
120 First during pruning First pruningofAt the time
121 The code at the top is wrong. The code on p126 is correct
145 5 and 10 5,10,20 pieces
148 Data here Individualdata
149 Numpy "agerage" function NumPy's "average"function
149~ WeighedZeroRule WeightedZeroRule
151~ WeighedDecisionStump WeightedDecisionStump
163 (Equation 3)ofyi y_i
164 (Equation 8), First minus
166 Class variables Instance variables
166 Overload override
167~ itre iter
174 Add up the weights for the correct data Incorrect answerAdd weights to data
175 The total weight of the correct data Sum of the weights of the correct dataso
190 At the time when it becomes half of all the total Half of all totalTaAt the time
200 First time from AdaBoost From AdaBoostStart with
200 So that the value decreases Its value decreaseslike
200 There is no one-character space for changing paragraphs on the last line
211~ ganma gamma

Recommended Posts

Collect typographical information from Toshiyuki Sakamoto's "Make and Understand! Introduction to Ensemble Learning Algorithms"
[Introduction to Tensorflow] Understand Tensorflow properly and try to make a model
Introduction to Deep Learning ~ Convolution and Pooling ~
Machine learning algorithms (from two-class classification to multi-class classification)
Introduction to Deep Learning ~ Localization and Loss Function ~
Artificial intelligence, machine learning, deep learning to implement and understand
An introduction to machine learning from a simple perceptron