fork download
copy 
  1. import java.util.HashMap;
  2. import java.util.HashSet;
  3. import java.util.Map;
  4. import java.util.Set;
  5. import java.util.stream.Collectors;
  6.  
  7. class Test
  8. {
  9.     private static Map<String, Integer> tagToIgnores = new HashMap<>();
  10.     private static Map<String, Integer> tagToUsage = new HashMap<>();
  11.  
  12.     static
  13.     {
  14.         tagToIgnores.put("c#", 23790);
  15.         tagToIgnores.put("php", 23721);
  16.         tagToIgnores.put("java", 22250);
  17.         tagToIgnores.put("android", 22201);
  18.         tagToIgnores.put("asp.net", 17320);
  19.         tagToIgnores.put("c++", 15191);
  20.         tagToIgnores.put("python", 14869);
  21.         tagToIgnores.put(".net", 13293);
  22.         tagToIgnores.put("ruby", 12691);
  23.         tagToIgnores.put("ios", 11896);
  24.         tagToIgnores.put("ruby-on-rails", 11819);
  25.         tagToIgnores.put("iphone", 10955);
  26.         tagToIgnores.put("javascript", 10078);
  27.         tagToIgnores.put("objective-c", 8587);
  28.         tagToIgnores.put("vb.net", 7892);
  29.         tagToIgnores.put("c", 7767);
  30.         tagToIgnores.put("jquery", 6939);
  31.         tagToIgnores.put("delphi", 6563);
  32.         tagToIgnores.put("django", 6423);
  33.         tagToIgnores.put("perl", 5925);
  34.         tagToIgnores.put("asp.net-mvc", 5887);
  35.         tagToIgnores.put("flash", 5569);
  36.         tagToIgnores.put("mysql", 5142);
  37.         tagToIgnores.put("windows", 5131);
  38.         tagToIgnores.put("facebook", 4851);
  39.         tagToIgnores.put("r", 4629);
  40.         tagToIgnores.put("wpf", 4484);
  41.         tagToIgnores.put("visual-studio", 4392);
  42.         tagToIgnores.put("angularjs", 4298);
  43.         tagToIgnores.put("flex", 4186);
  44.         tagToIgnores.put("sql-server", 4108);
  45.         tagToIgnores.put("wordpress", 4076);
  46.         tagToIgnores.put("silverlight", 4016);
  47.         tagToIgnores.put("oracle", 4005);
  48.         tagToIgnores.put("css", 3972);
  49.         tagToIgnores.put("haskell", 3752);
  50.         tagToIgnores.put("scala", 3737);
  51.         tagToIgnores.put("html", 3663);
  52.         tagToIgnores.put("vba", 3624);
  53.         tagToIgnores.put("sql", 3359);
  54.         tagToIgnores.put("node.js", 3232);
  55.         tagToIgnores.put("matlab", 3220);
  56.         tagToIgnores.put("sharepoint", 3108);
  57.         tagToIgnores.put("eclipse", 3073);
  58.         tagToIgnores.put("xcode", 3068);
  59.         tagToIgnores.put("swift", 3065);
  60.         tagToIgnores.put("angular", 2956);
  61.         tagToIgnores.put("spring", 2890);
  62.         tagToIgnores.put("linux", 2875);
  63.         tagToIgnores.put("macos", 2804);
  64.  
  65.         tagToUsage.put("c#", 1286491);
  66.         tagToUsage.put("php", 1263058);
  67.         tagToUsage.put("java", 1515802);
  68.         tagToUsage.put("android", 1173706);
  69.         tagToUsage.put("asp.net", 342947);
  70.         tagToUsage.put("c++", 605528);
  71.         tagToUsage.put("python", 1114060);
  72.         tagToUsage.put(".net", 279901);
  73.         tagToUsage.put("ruby", 202463);
  74.         tagToUsage.put("ios", 591066);
  75.         tagToUsage.put("ruby-on-rails", 303176);
  76.         tagToUsage.put("iphone", 219814);
  77.         tagToUsage.put("javascript", 1763733);
  78.         tagToUsage.put("objective-c", 286812);
  79.         tagToUsage.put("vb.net", 122490);
  80.         tagToUsage.put("c", 296656);
  81.         tagToUsage.put("jquery", 944680);
  82.         tagToUsage.put("delphi", 43337);
  83.         tagToUsage.put("django", 190837);
  84.         tagToUsage.put("perl", 61725);
  85.         tagToUsage.put("asp.net-mvc", 178208);
  86.         tagToUsage.put("flash", 35282);
  87.         tagToUsage.put("mysql", 550486);
  88.         tagToUsage.put("windows", 127560);
  89.         tagToUsage.put("facebook", 83210);
  90.         tagToUsage.put("r", 276465);
  91.         tagToUsage.put("wpf", 147423);
  92.         tagToUsage.put("visual-studio", 85729);
  93.         tagToUsage.put("angularjs", 257148);
  94.         tagToUsage.put("flex", 24297);
  95.         tagToUsage.put("sql-server", 253505);
  96.         tagToUsage.put("wordpress", 145586);
  97.         tagToUsage.put("silverlight", 27515);
  98.         tagToUsage.put("oracle", 106795);
  99.         tagToUsage.put("css", 574065);
  100.         tagToUsage.put("haskell", 39976);
  101.         tagToUsage.put("scala", 86886);
  102.         tagToUsage.put("html", 804877);
  103.         tagToUsage.put("vba", 139729);
  104.         tagToUsage.put("sql", 479302);
  105.         tagToUsage.put("node.js", 262844);
  106.         tagToUsage.put("matlab", 82398);
  107.         tagToUsage.put("sharepoint", 24236);
  108.         tagToUsage.put("eclipse", 115751);
  109.         tagToUsage.put("xcode", 130502);
  110.         tagToUsage.put("swift", 221958);
  111.         tagToUsage.put("angular", 153902);
  112.         tagToUsage.put("spring", 147214);
  113.         tagToUsage.put("linux", 173188);
  114.         tagToUsage.put("macos", 90234);
  115.     }
  116.  
  117.     public static void main(String[] args)
  118.     {
  119.         final double averageFreq = tagToIgnores.values().stream().mapToInt(x -> x).average().orElse(0);
  120.         final double averageUsage = tagToUsage.values().stream().mapToInt(x -> x).average().orElse(0);
  121.  
  122.         final Map<String, Double> tagToNormalisedIgnores = tagToIgnores.entrySet()
  123.             .stream()
  124.             .collect(Collectors.toMap(Map.Entry::getKey, v -> v.getValue() / averageFreq));
  125.  
  126.         final Map<String, Double> tagToNormalisedUsage = tagToUsage.entrySet()
  127.             .stream()
  128.             .collect(Collectors.toMap(Map.Entry::getKey, v -> v.getValue() / averageUsage));
  129.  
  130.         final double similarity = cosineSimilarity(tagToNormalisedIgnores, tagToNormalisedUsage);
  131.         System.out.println(similarity);
  132.         System.out.println();
  133.  
  134.  
  135.         final Map<String, Double> tagToDiff = new HashMap<>();
  136.         for (final String tag : tagToNormalisedIgnores.keySet())
  137.         {
  138.             double diff = tagToNormalisedIgnores.get(tag) - tagToNormalisedUsage.get(tag);
  139.             tagToDiff.put(tag, diff);
  140.         }
  141.         tagToDiff.entrySet().stream()
  142.             .sorted(Map.Entry.comparingByValue())
  143.             .forEach(System.out::println);
  144.     }
  145.  
  146.     private static double cosineSimilarity(Map<String, Double> a, Map<String, Double> b)
  147.     {
  148.         final Set<String> intersection = new HashSet<>(a.keySet());
  149.         intersection.retainAll(b.keySet());
  150.  
  151.         double dotProduct = 0, magnitudeA = 0, magnitudeB = 0;
  152.  
  153.         for (String item : intersection) {
  154.             dotProduct += a.get(item) * b.get(item);
  155.         }
  156.         for (String k : a.keySet()) {
  157.             magnitudeA += Math.pow(a.get(k), 2);
  158.         }
  159.         for (String k : b.keySet()) {
  160.             magnitudeB += Math.pow(b.get(k), 2);
  161.         }
  162.         return dotProduct / Math.sqrt(magnitudeA * magnitudeB);
  163.     }
  164. }
Success #stdin #stdout 0.18s 2249728KB
comments (?)
stdin
copy 
Standard input is empty
stdout
copy 
0.8491376330293606

javascript=-3.4483760137962687
html=-1.6976975129079275
jquery=-1.6427959311647533
java=-1.1630527929353662
python=-1.0514340363884729
css=-1.0307223131018295
sql=-0.8549228736433856
mysql=-0.8117372652480911
c#=-0.337034124217642
node.js=-0.28465252302870464
php=-0.28261974629541253
android=-0.2416793029393749
swift=-0.19588597674892733
sql-server=-0.14324730015057452
r=-0.1364861438430588
angularjs=-0.1279521013253736
linux=-0.08878330674595541
ios=-0.026858206316774602
angular=-0.025740959776942463
spring=-0.016346493373397386
xcode=0.052567460394544796
eclipse=0.09323909615860027
vba=0.1012143068852851
macos=0.126804744872855
wordpress=0.14522099519834297
wpf=0.19430086642043298
matlab=0.20318072032609627
c=0.22455301918775117
oracle=0.24102583319651388
scala=0.2595134208923015
asp.net-mvc=0.2967083986249241
django=0.33347789173710907
windows=0.333906486721583
sharepoint=0.34609274417622427
visual-studio=0.3494428423804226
objective-c=0.3599076282539395
c++=0.37052312068189175
haskell=0.38873490379278347
facebook=0.41709527871486446
silverlight=0.45751437703026926
flex=0.4887685361645861
perl=0.6176802188083449
flash=0.6422294516349092
vb.net=0.7135063248214386
ruby-on-rails=0.7437822997821084
delphi=0.7520926475436067
iphone=0.8553998770887196
.net=1.0022245693771021
ruby=1.1324910108396733
asp.net=1.3648258522410086
https://ideone.com/PcVHMa
language:
Java (HotSpot 12)
created:
5 years ago
visibility:
public

Share or Embed source code

Discover > Sphere Engine API

The brand new service which powers Ideone!

Discover > IDE Widget

Widget for compiling and running the source code in a web browser!