Scribd Logo
Upload

Welcome to Scribd!

  • 7 views

    Compute2

    Uploaded by

    lil Aady
    The document discusses performing different clustering algorithms on a dataset including K-means clustering for various values of K, fuzzy C-means clustering, bottom-up clustering using agglomerative clustering and generating dendrograms, and density-based clustering using DBSCAN. Clusters from each method are captured and compared in a summary.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    Compute2

    Uploaded by

    lil Aady
    7 views10 pages
    The document discusses performing different clustering algorithms on a dataset including K-means clustering for various values of K, fuzzy C-means clustering, bottom-up clustering using agglomerative clustering and generating dendrograms, and density-based clustering using DBSCAN. Clusters from each method are captured and compared in a summary.

    Original Title

    12112120_compute2

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses performing different clustering algorithms on a dataset including K-means clustering for various values of K, fuzzy C-means clustering, bottom-up clustering using agglomerative clustering and generating dendrograms, and density-based clustering using DBSCAN. Clusters from each method are captured and compared in a summary.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    7 views10 pages

    Compute2

    Uploaded by

    lil Aady
    The document discusses performing different clustering algorithms on a dataset including K-means clustering for various values of K, fuzzy C-means clustering, bottom-up clustering using agglomerative clustering and generating dendrograms, and density-based clustering using DBSCAN. Clusters from each method are captured and compared in a summary.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    You are on page 1of 10

    1st task: perform data cleaning, if any, in the dataset.

    As the data is already cleaned so there is no requirement of data cleaning

    import numpy as np
    import pandas as pd
    import os

    x = pd.read_csv('data.csv')

    x.dropna()
    print(x)

    2nd task: perform K-means Clustering for K=3,5,7 and also Fuzzy C means. Capture the Clusters
    generated with Both K Means & C means.

    import pandas as pd
    from sklearn.cluster import KMeans

    from sklearn.preprocessing import StandardScaler


    from sklearn.decomposition import PCA
    from fcmeans import FCM

    import matplotlib.pyplot as plt

    z = pd.read_csv('data.csv')

    X = z.iloc[:, 1:59].values

    scale = StandardScaler()
    X = scale.fit_transform(X)

    pca = PCA(n_components=2)
    X_pca = pca.fit_transform(X)

    k_values = [3, 5, 7]
    fuzzy_cmeans_c = [3, 5, 7]

    for k in k_values:
    kmeans = KMeans(n_clusters=k, random_state=42)
    y_kmeans = kmeans.fit_predict(X_pca)

    plt.figure(figsize=(6, 4))
    plt.scatter(X_pca[:, 0], X_pca[:, 1], c=y_kmeans, cmap='viridis')
    plt.title('K-means clustering (K = ' + str(k) + ')')
    plt.xlabel('Principal Component 1')
    plt.ylabel('Principal Component 2')
    plt.show()
    for c in fuzzy_cmeans_c:
    fcm = FCM(n_clusters=10, m=c)
    fcm.fit(X_pca)
    y_fcm = fcm.predict(X_pca)

    plt.figure(figsize=(6, 4))
    plt.scatter(X_pca[:, 0], X_pca[:, 1], c=y_fcm, cmap='viridis')
    plt.title('Fuzzy C means clustering (c = ' + str(c) + ')')
    plt.xlabel('Principal Component 1')
    plt.ylabel('Principal Component 2')
    plt.show()

    OUTPUT
    3rd task: perform Bottom-up Clustering (Agglomerative clustering). Capture the Clusters generated
    at a different level, and also prepare dendrograms.

    import pandas as pd
    import numpy as np
    import matplotlib.pyplot as plt
    from sklearn.cluster import AgglomerativeClustering
    from scipy.cluster.hierarchy import dendrogram, linkage

    Data = pd.read_csv('data.csv')

    mfs = np.array(Data.iloc[:, 1:59])

    agg_clustering = AgglomerativeClustering(n_clusters=None, linkage='ward',


    distance_threshold=0)
    cluster = agg_clustering.fit_predict(mfs)

    linked = linkage(mfs, method='ward')


    dendrogram(linked, truncate_mode='lastp', p=30, orientation='top')
    plt.show()
    for i in range(2, 12):
    clustering = AgglomerativeClustering(n_clusters=i, linkage='ward')
    clustering.fit(mfs)
    print(f'Clusters at level {i}: {clustering.labels_}')

    # Observations:

    # The dendogram shows hierarchy of clusters formed by the agglomerative


    clustering algorithm
    # The clusters start merging from the bottom level and go up to the top level.
    # At the top level, we can see that all the data points belong to a single
    cluster.
    # By looking at the dendrogram, we can choose the appropriate level to get the
    desired number of clusters.
    # We can also see that at each level, the clustering algorithm forms a
    different set of clusters based on the distance threshold and linkage
    criterion used.
    4rth task: perform density-based (DBSCAN) Clustering,
    5th Task: prepare a brief Comparative summary of clusters generated using the above clustering
    techniques.

    import pandas as pd
    import numpy as np
    import matplotlib.pyplot as plt
    from sklearn.cluster import DBSCAN
    from sklearn.decomposition import PCA

    data = pd.read_csv('data.csv')

    mfs = np.array(data.iloc[:, 1:-1])

    dbscn_clustering = DBSCAN(eps=20, min_samples=5)


    clusters = dbscn_clustering.fit_predict(mfs)

    pc = PCA(n_components=2)
    reduced_features = pc.fit_transform(mfs)

    plt.scatter(reduced_features[:, 0], reduced_features[:, 1], c=clusters,


    cmap='viridis')
    plt.show()

    # Observations:
    # DBSCAN clustering algorithm forms clusters based on the density of the data
    points.
    # The resulting clusters are not well

    You might also like