ML Practice 6_3
Dimensionaliy Reduction: Decreasing the size of the data by selecting some features that best represent the data To prevent overfitting and improve model performance PCA(Principal Component Analysis),
Dimensionaliy Reduction: Decreasing the size of the data by selecting some features that best represent the data To prevent overfitting and improve model performance PCA(Principal Component Analysis),
K-means Clustering Find mean of pixel value : cluster center, centroid Determine the centers of k clusters at random. Find the nearest cluster center from each sample and designate it as a sample of t
Unsupervised Learning No dependent variables and targets. (↔ Supervised Learning) Clustering (Multiple class) Must be many different types of data Linked to deep learning Dimensionality reduction I
Goal : To change the color of tree plot1!pip install -U matplotlib Requirement already satisfied: matplotlib in c:\programdata\anaconda3\lib\site-packages (3.4.3) Collecting matplotlib Downloading
Ensemble algorithm that performs best in dealing with structured data Bagging : A method of aggregating results by taking multiple bootstrap samples and training each model. (parallel learning) Random
Cross Validation: Repeated process of spliting validation set and evaluating model. Train set : Validation set : Test set = 6 : 2 : 2 (generally) Test sets are not used in the model learning proc
Prepare Data Import wine data set class 0: red wine class 1: white wine 123import pandas as pdwine = pd.read_csv("https://bit.ly/wine_csv_data")print(wine.head()) alcohol sugar pH
Gradient Descent(경사 하강법): Algorithm for finding the minimum value of a loss function using a sample of a training set stochastic gradient descent(확률적 경사 하강법; SGD) method of randomly selecting one samp
Prepare DataImport data set1234import pandas as pdfish = pd.read_csv('https://bit.ly/fish_csv_data')print(fish.head()) Species Weight Length Diagonal Height Width 0 Bream 242.0
Prepare Data123import pandas as pddf = pd.read_csv('https://bit.ly/perch_csv_data')perch_full = df.to_numpy() # Convert Pandas DataFrame to Numpy Array 12345678import numpy as npperch_weig