PPDB-Recommendation-Algorithms

A library of recommendation algorithms for the assignment of "Programming Project Databases 2020-2021".

Contents

The folder src/algorithm contain 5 recommendation algorithms:

• Random
• Popularity
• Item Nearest Neighbours (cosine)
• EASE
• Weighted Matrix Factorization (WMF)

You can test them with the provided CLI in main.py.

Additionally in the file src/util.py there are common utility functions for manipulating csv files, numpy arrays and scipy sparse arrays for recommendation.

Example Usage (CLI)

>python3 main.py iknn ./data/interactions.csv --k 100 --normalize
User 8370
History [   5    6   11   15   19   23   43   54   95   96  102  115  146  148
  151  154  156  159  166  182  185  188  196  199  201  207  213  219
  220  221  248  265  270  272  312  318  329  340  341  356  357  358
  360  367  380  417  466  476  495  496  560  569  577  578  580  582
  626  628  670  671  672  678  710  727  795  797  951  953 1013 1026
 1064 1189 1249 1403 1586 1680 1839 2069 2329 3498 3934]
recommendations [  5 188   6 185  37]
scores [0.75590206 0.72380511 0.71297999 0.70977589 0.68067831]

User 11748
History [  57   74   75   90   96  121  123  125  127  129  130  132  138  140
  141  142  143  155  202  223  226  228  232  234  236  237  241  245
  251  268  275  310  355  356  380  382  395  400  410  412  416  433
  434  437  438  442  444  445  463  488  543  572  574  591  595  600
  601  622  634  642  656  663  670  676  699  730  733  736  760  764
  765  775  875  938  947  992  995 1000 1025 1051 1053 1060 1075 1079
 1084 1085 1087 1090 1091 1093 1095 1096 1099 1104 1118 1120 1127 1130
 1163 1166 1169 1171 1214 1248 1276 1315 1317 1319 1321 1324 1325 1326
 1351 1365 1394 1425 1441 1455 1475 1522 1523 1525 1527 1528 1530 1563
 1572 1578 1583 1608 1614 1619 1636 1641 1647 1662 1694 1697 1730 1739
 1740 1814 1816 1819 1822 1833 1834 1837 1845 1851 1866 1915 1920 1925
 2036 2041 2052 2098 2107 2117 2119 2141 2216 2218 2235 2241 2260 2276
 2281 2318 2323 2344 2362 2403 2476 2483 2488 2490 2504 2513 2566 2583
 2600 2666 2685 2796 2802 2957 3046 3057 3063 3086 3093 3140 3165 3318
 3359 3377 3386 3399 3451 3496 3497 3521 3526 3548 3550 3553 3571 3577
 3605 3638 3657 3714 3732 3733 3798 3927 3941 3947 4006 4013 4027 4029
 4041 4043 4100 4117 4127 4142 4162 4267]
recommendations [434 137 596 135 228]
scores [0.93871955 0.88715344 0.83393249 0.81350513 0.80768068]

User 17368
History [ 154  172  176  236  287  306  330  331  332  374  397  411  437  465
  512  524  542  544  556  572  602  636  660  692  710  714  727  755
  788  789  798  831  914 1157 1259 1264 1267 1287 1297 1350 1427 1476
 1491 1515 1702 1874 1878 2103 2160 2199 2207 2352 2404 2425 2542 2597
 2662 2738 2746 2754 2792 3120 3123 3541 4024]
recommendations [466 191 174 166 376]
scores [0.28427849 0.25856493 0.25823504 0.25039131 0.2399928 ]

User 950
History [   1    3    4    5   20   23   37   54   65  100  146  147  148  149
  154  156  166  173  174  175  179  182  185  186  188  189  190  193
  198  200  201  208  216  226  249  252  253  258  264  270  272  273
  274  283  286  289  300  311  340  341  349  355  358  363  364  368
  374  383  405  418  447  455  465  466  481  488  490  491  495  508
  512  515  520  526  557  578  604  605  612  622  655  658  675  679
  706  719  724  730  736  740  741  744  791  805  816  830  853  870
  934 1011 1082 1110 1144 1182 1263 1289 1303 1499 1506 1555 1588 1714
 1953 2087 2143 2224 2244 2248 2265 2531 2583 2704 2707 2831 3120 3222
 3719 3759]
recommendations [146 208   6   5 148]
scores [1.14039124 1.04338577 1.03027084 1.01427813 1.00576703]

User 15257
History [   6   95  155  189  221  228  234  247  248  253  281  337  340  368
  461  464  658  885 1285 1324 1619 1730 4191]
recommendations [146 208   6 173 184]
scores [0.2189981  0.21721212 0.20357215 0.19742866 0.19552239]