Clustering Analysis Algorithms

Implementations

• K-Means Algorithm
• DB-Scan Density based apatial clustering of applications with Noise
• EM-Algo Expection Maximization Algorithm
• Denclue Density based clustering

Dataset Links

• Iris Dataset
• Spiral Dataset

RUN Commands

K-Means_DBscan.cpp

Input

---

• Step-1. keep The input datasetfile within the same folder and set the file name in takeinputdataset() function (on the top of the code).
• step-2. choose the value of the number of cluster and set on the top of the code #define k
• step-3. set maxDBesp (range for the density based clustering) value on the top(#define maxDBeps).
• Run the K-Means_DBscan.cpp file in c++ compiler.

output

---

• you will get Final Expectations(mean) with the total number of clusters .
• It will also generate one output.txt file in the same directory having all the clusters with their corresponding datapoints.

EM-Algo_Denclue.py

• Note Please use the jupyter notebook for better visualisation of the graph and to support all the matplotlib module in the code.

Input:

---

• Step1. keep The input datasetfile within the same folder and set the file name in file1 = open("irisdataset.txt") (on the top of the code).
• step2. set the Dimension as you want (by default dim = 2 on the top of the code) Run file EM-Algo_Denclue.py

Output:

---

• you will get Dataset Estimated parameters Mean , Covariance Matrices and Probability for each cluster in your console.
• Than it will ask you to show the graph. for EM-Algo press-1 for Denclue Press-2 press-1/2 (correspondingly) you will get the graph in your console(for that use jupyter notebook )

Refer DetailedReport.txt for the detailed analysis and comparison of the algorithms over particular shape distribution dataset.