Detailed description of the k-means clustering algorithm implemented by Java, k-means clustering

Detailed description of the k-means clustering algorithm implemented by Java, k-means clustering

Requirement

Execute the k-means algorithm for a field in a table in the MySQL database to write the processed data to the new table.

Source code and driver

Kmeans_jb51.rar

Source code

Import java. SQL. *; import java. util. *;/*** @ author tianshl * @ version 2018/1/13 am */public class Kmeans {// source data private List <Integer> origins = new ArrayList <> (); // private Map of grouped data <Double, List <Integer> grouped; // The initial centroid List private List <Double> cores; // the data source private String tableName; private String colName; /*** constructor ** @ param tableName source data table name * @ param colName source data column name * @ param cores center list * /Private Kmeans (String tableName, String colName, List <Double> cores) {this. cores = cores; this. tableName = tableName; this. colName = colName;}/*** recalculate the centroid ** @ return New centroid List */private List <Double> newCores () {List <Double> newCores = new ArrayList <> (); for (List <Integer> v: grouped. values () {newCores. add (v. stream (). reduce (0, (sum, num)-> sum + num)/(v. size () + 0.0);} Collections. sort (newCo Res); return newCores;}/*** determine whether to end ** @ return bool */private Boolean isOver () {List <Double> _ cores = newCores (); for (int I = 0, len = cores. size (); I <len; I ++) {if (! Cores. get (I ). toString (). equals (_ cores. get (I ). toString () {// use the new centroid cores = _ cores; return false;} return true;}/*** Data Group */private void setGrouped () {grouped = new HashMap <> (); Double core; for (Integer origin: origins) {core = getCore (origin); if (! Grouped. containsKey (core) {grouped. put (core, new ArrayList <> ();} grouped. get (core ). add (origin) ;}}/*** select the center of mass ** @ param num the data to be grouped * @ return center of mass */private Double getCore (Integer num) {// difference List <Double> diffs = new ArrayList <> (); // calculate the difference for (Double core: cores) {diffs. add (Math. abs (num-core);} // Minimum deviation-> index-> return cores of the center of mass. get (diffs. indexOf (Collections. min (diffs);}/*** create a database connection * @ Return connection */private Connection getConn () {try {// URL points to the name of the database to be accessed mydata String url = "jdbc: mysql: // localhost: 3306/data_analysis_dev "; // MySQL username String user = "root"; // MySQL password String password = "root"; // load the driver Class. forName ("com. mysql. jdbc. driver "); // declare the Connection object Connection conn = DriverManager. getConnection (url, user, password); if (conn. isClosed () {System. out. println ("Connections Failed to database! "); Return null;} System. out. println (" database connected successfully! "); Return conn;} catch (Exception e) {System. out. println (" failed to connect to the database! "); E. printStackTrace ();} return null;}/*** close database Connection ** @ param conn Connection */private void close (Connection conn) {try {if (conn! = Null &&! Conn. isClosed () conn. close ();} catch (Exception e) {e. printStackTrace () ;}}/*** obtain source data */private void getOrigins () {Connection conn = null; try {conn = getConn (); if (conn = null) return; Statement statement = conn. createStatement (); ResultSet rs = statement.exe cuteQuery (String. format ("select % s from % s", colName, tableName); while (rs. next () {origins. add (rs. getInt (1);} conn. close ();} cat Ch (Exception e) {e. printStackTrace () ;}finally {close (conn) ;}/ *** write data to the new table */private void write () {Connection conn = null; try {conn = getConn (); if (conn = null) return; // Create a table Statement statement = conn. createStatement (); // Delete the old data TABLE statement.exe cute ("drop table if exists k_means ;"); // CREATE a new TABLE statement.exe cute ("create table if not exists k_means ('core' DECIMAL (11, 7), 'col' INTEGER (11 )); "); // Disable automatic submission of conn. setAutoCommit (false); PreparedStatement ps = conn. prepareStatement (" insert into k_means VALUES (?, ?) "); For (Map. entry <Double, List <Integer> entry: grouped. entrySet () {Double core = entry. getKey (); for (Integer value: entry. getValue () {ps. setDouble (1, core); ps. setInt (2, value); ps. addBatch () ;}/// Batch Execute ps.exe cuteBatch (); // submit the transaction conn. commit (); // close the connection conn. close ();} catch (Exception e) {e. printStackTrace () ;}finally {close (conn) ;}/ *** process data */private void run () {System. out. println ("Get Source data "); // get source data getOrigins (); // stop grouping Boolean isOver = false; System. out. println (" data group processing "); while (! IsOver) {// data group setGrouped (); // determine whether to stop the group isOver = isOver ();} System. out. println ("write processed data to the database"); // write grouped data to the new table write (); System. out. println ("Data Writing completed");} public static void main (String [] args) {List <Double> cores = new ArrayList <> (); cores. add (1, 260.0); cores. add (600.0); // table name, column name, centroid list new Kmeans ("attributes", "attr_length", cores ). run ();}}

Source File

Kmeans.java

Compile

javac Kmeans.java 

Run

# Specify the dependent library java-Djava. ext. dirs =./lib Kmeans

The above is all the content of this article. I hope it will be helpful for your learning and support for helping customers.