スキップしてメイン コンテンツに移動

Implementation of Logistic Regression

Introduction


Today, I implement Logistic Regression.
My OS of computer is the windows10.
Implementation is used by Python3.
I use the IRLS to estimate optimization value.
I introduce the theory of Logistic Regression in another post.
If you interested, look at this post.

Overview

  • I will introduce used data set
  • I will introduce my code in Python
  • I will show you result on Command line.

Dataset

I use this dataset to implement Logistic Regression.
This dataset is Residential area data.
I diplay this data in Pandas DataFrame Python3.
enter image description here
This is data set from top to five elements.
if people live the house, occupancy is 1.
if people do not live the house, occupancy is 0.
This data consist of 8000 samples to use as training data, and 2000 samples to use as test data.
However I use 100 samples as training data and 100 samples as test data, because my computer is not designated programing.
Sorry, .

CODE

This code is very long.
Thus, I publish my code of Logistice Regression in my Github.
My github page
My Logistic Regression code(def file)
My Logistic Regression code(main file)
Separating my code have reason. It is that I want to separate define file and main file. main file have
if __name__ == '__mian__'
my def file have algorithm of Logistic Regression. My code is defined class in def file. I will write about class in Python.

Execution!

w is estimating…
enter image description here
I will save figure of value of Closs-entropy error function
enter image description here
This is sactter plot of closs entropy error function.
enter image description here
I find out decreasing of value of closs entropy error funtion.
I finished estimating optimization.
I will test my model of Logistic Regression.
enter image description here
enter image description here
I compare my predict by Logistic Regression and correct class.
Percentage of correct answer is 98per.
I think this is high score.
By the way, Logistic Regression find out probability that each data point exists \(C_1\)
Please check out P columns.
As long as I identify, Almost the P is not near 0.5.
Labels:

コメント

コメントを投稿

このブログの人気の投稿

Implementation of Robbins monro

Robbins monro の実装 sorry, this page is Japanese only.   今回はRobbins monro の実装をしてみました。 Robbins monroは確率勾配降下法の学習率を入りテーション回数の逆数で割っていくものです。 使っているprogram言語はpython 3です。osはwindowsです。(macほしい...) アルゴリズム 確率勾配降下方とは目的関数の最適解を求めるアルゴリズムです。目的関数をf(X)とすると、手順は以下のようになっています。 初期学習率$n_0$を決めます。訓練データDを用意します。この訓練データは複数の初期値の集まりです。 訓練データから一つ初期値をランダムに取り出し、これを$x_0$とし、最初の予測値とします。 次の式に現在の予測値$x_0$を代入し、新たな予測値$x_{n+1}$を得ます。$$x_{n+1} = x_{n} - \frac{n_0}{n} grad f(X_n)$$ 収束して入れば4へ、収束していなければ2で得られた値$x{n+1}$を新たに$x_n$としてもう一度2を行う。 訓練データを一周していなければ2へ、一周していれば各初期値から得られた解の中から目的関数を最も小さくするものを選ぶ。　　 実装例 以下の目的関数を最小化させてみましょう。 $$f(x,y) = (x-2)^2 + (y-3)^2 $$ コマンドラインでpythonを実行していきます。 予想通り、（２，３）という解を導き出してくれました。目的関数が簡単だったので、初期値をどの値でとってもばっちり正解にたどり着いてくれました。 CODE 以下にRobbins monroの関数だけ置いておきます。 こちら にすべてのコードを載せています。 def Robbins_monro(function,grad,number_variable_gradient): init_learning_rate = 1.5 stepsize = 1000 init_value = np.array([range(-1000,1020,20) for i in range(number_v...
コメントを投稿
続きを読む

Graph theory in Python

Introduction 日本語 ver Today, I write the networkx module for graph theory in Python. I write about basis of Gragh theory. Please look at this page . This post is not written Graph theory, but I want to share module in Python for Graph theory. Networkx Python3 have Networkx module. If you have python2, you should install by pip pip install networkx Make Graph by Networkx first, import networkx module and make instance. import networkx as nx import matplotlib.pyplot as plt G = nx.Graph() next, add node and edge G.add_node(1) # add Multiple nodes G.add_nodes_from([2,3,4]) G.add_edge(1,2) # add Multiple edges G.add_edges_from([(3,4),(1,2),(4,6)]) next, plot G nx.draw(G) plt.show() networkx have many function. I will write new learning about networkx in this post. Reference https://qiita.com/kzm4269/items/081ff2fdb8a6b0a6112f http://akiniwa.hatenablog.jp/entry/2013/05/12/012459
コメントを投稿
続きを読む

LIMIT and user variable in MysQL

Problem 日本語 ver I want to use this query in PROCEDURE. SET @user_variable = FLOOR(RAND()*10); INSERT INTO  Table_name (columns_name) VALUES (1) FROM Table_name ORDER BY RAND() LIMIT 1 OFFSET @user_variable; This query want to substitute 1 for the @user_variable 'th line of columns_name, but this query does not work. A cause is that LIMIT and OFFSET does not user variable. Solution SET @user_variable = FLOOR(RAND()*10); PREPARE SET_STMT FROM 'INSERT INTO  Table_name (columns_name) VALUES (1) FROM Table_name ORDER BY RAND() LIMIT 1 OFFSET ?;';  EXECUTE SET_STMT USING @user_variable; Conclusion, if you want to use user variable with LIMIT or OFFSET, use PREPARE STATMENT. Reference http://techtipshoge.blogspot.com/2011/10/limit.html
2 件のコメント
続きを読む