diff --git a/Experiments.ipynb b/Experiments.ipynb
index 841f82ca1ebe10cb41b12499cdb28b776bb571f2..20bad0e0e28e46e6e5bcb92a18cee7f141ddd5fa 100644
--- a/Experiments.ipynb
+++ b/Experiments.ipynb
@@ -1,17 +1,5 @@
{
"cells": [
- {
- "cell_type": "markdown",
- "source": [
- "# Experiment notebook\n",
- "\n",
- "The following jupyter notebook contains all the commands used reproduce the experiments of the paper"
- ],
- "metadata": {
- "collapsed": false
- },
- "id": "523fb0ddcbd44033"
- },
{
"cell_type": "code",
"outputs": [],
@@ -20,145 +8,51 @@
"datasets = ['assist09_tkde', 'assist17_tkde', 'algebra','math_1', 'math_2']"
],
"metadata": {
- "ExecuteTime": {
- "end_time": "2024-02-14T09:17:51.136290823Z",
- "start_time": "2024-02-14T09:17:51.130011719Z"
- }
+ "collapsed": false
},
- "id": "9d54cd9a",
- "execution_count": 3
+ "id": "b0ab105c19fde634"
},
{
"cell_type": "markdown",
- "id": "1da92e4f",
- "metadata": {},
"source": [
- "## Table 2: compute ACC, AUC and RMSE\n",
+ "## Table 2 and 3: compute ACC, AUC and RMSE and DOA\n",
"\n",
"Warning : can be long to compute\n",
"Prerequisite : Running the makefile\n",
"Output : stored in `../../results/table_2`"
- ]
+ ],
+ "metadata": {
+ "collapsed": false
+ },
+ "id": "841ae3856df97ef0"
},
{
"cell_type": "code",
- "execution_count": 7,
- "id": "61c53cb1",
- "metadata": {
- "ExecuteTime": {
- "end_time": "2024-02-14T09:39:50.059022516Z",
- "start_time": "2024-02-14T09:39:12.027137257Z"
- }
- },
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "/home/arthurb/Programmation/cd-bpr\n",
- "assist0910_tkde\r\n",
- "Is CUDA supported by this system? False\r\n",
- "CUDA version: None\r\n",
- "==========> fold 0\r\n",
- "dataTrain: ../../data/assist0910_tkde/train_valid_0.csv\r\n",
- "dataTest: ../../data/assist0910_tkde/test_0.csv\r\n",
- "embPath: ../../results/table_2/\r\n",
- "epochs: 1\r\n",
- "batch_size: 512\r\n",
- "[Epoch 0] loss: 1.370641\r\n",
- "Doa: 0.7126622039356039\r\n",
- "AUC and RMSE: 0.7249253060269787 0.4473763215794524\r\n",
- "==========> fold 1\r\n",
- "dataTrain: ../../data/assist0910_tkde/train_valid_1.csv\r\n",
- "dataTest: ../../data/assist0910_tkde/test_1.csv\r\n",
- "embPath: ../../results/table_2/\r\n",
- "epochs: 1\r\n",
- "batch_size: 512\r\n",
- "^C\r\n",
- "Traceback (most recent call last):\r\n",
- " File \"/home/arthurb/Programmation/cd-bpr/code/binary_bpr/main.py\", line 312, in <module>\r\n",
- " dico_items, test, y_test = parse_dataframe(dataTest, dico_kc, dico_users, dico_items, False)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/Programmation/cd-bpr/code/binary_bpr/main.py\", line 105, in parse_dataframe\r\n",
- " for row_index, row in df_group.iterrows():\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/frame.py\", line 1449, in iterrows\r\n",
- " for k, v in zip(self.index, self.values):\r\n",
- " ^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/frame.py\", line 12281, in values\r\n",
- " return self._mgr.as_array()\r\n",
- " ^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/internals/managers.py\", line 1656, in as_array\r\n",
- " arr = self._interleave(dtype=dtype, na_value=na_value)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/internals/managers.py\", line 1682, in _interleave\r\n",
- " dtype = interleaved_dtype( # type: ignore[assignment]\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/internals/base.py\", line 363, in interleaved_dtype\r\n",
- " return find_common_type(dtypes)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/dtypes/cast.py\", line 1428, in find_common_type\r\n",
- " types = list(dict.fromkeys(types).keys())\r\n",
- " ^^^^^^^^^^^^^^^^^^^^\r\n",
- "KeyboardInterrupt\r\n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"import os\n",
"\n",
"cmd = 'cd code/binary_bpr && python ./script.py '\n",
"!{cmd}"
- ]
+ ],
+ "metadata": {
+ "collapsed": false
+ },
+ "id": "f361888d67da2988"
},
{
"cell_type": "markdown",
- "id": "6b1c6b4d",
- "metadata": {},
"source": [
- "## Table 3: compute DOA\n",
- "\n",
- "The trained embeddings of CD-BPR and its competitors are given in the `result` directory so that the doa can be computed directly.\n",
- "Nonetheless, for the sake of reproducibility, they will be overwritten by the embeddings computed with the previous cell if you run it.\n",
- "\n",
- "After selecting the dataset with the variable i, the doa will be computed for each competitor and each data splitting of the cross validation.\n",
- "\n",
- "Warning : Can be long to compute\n",
- "Prerequisite : Running the makefile\n",
- "Output : displayed in the notebook"
- ]
+ "## Table 3: compute DOA"
+ ],
+ "metadata": {
+ "collapsed": false
+ },
+ "id": "d8116396a802bc7d"
},
{
"cell_type": "code",
- "execution_count": 17,
- "id": "790a43dd",
- "metadata": {
- "ExecuteTime": {
- "end_time": "2024-02-14T17:29:30.489316935Z",
- "start_time": "2024-02-14T17:15:40.682755850Z"
- }
- },
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "_create_directory() - handled error when creating a directory at ../experiments_logs/3: [Errno 17] File exists: '../experiments_logs/3'\r\n",
- "dataset math_1\r\n",
- "fold 0\r\n",
- "model DINA\r\n",
- "num_kc 11\r\n",
- "model MCD\r\n",
- "num_kc 11\r\n",
- "^C\r\n",
- "Traceback (most recent call last):\r\n",
- " File \"/home/arthurb/Programmation/cd-bpr/code/./DOA.py\", line 188, in <module>\r\n",
- " r_test = dao_creuse(F, kc_user, kc_user_val, dico_u)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/Programmation/cd-bpr/code/./DOA.py\", line -1, in dao_creuse\r\n",
- "KeyboardInterrupt\r\n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"import os\n",
"embDirPath = \"../../results/table_2/users/\"\n",
@@ -170,7 +64,11 @@
"\n",
"#doa = compute_doa(path+datasets[i]+'/train_embed.csv')\n",
"#print(\"DOA:\", doa)"
- ]
+ ],
+ "metadata": {
+ "collapsed": false
+ },
+ "id": "68a2dc1287b5cac"
},
{
"cell_type": "markdown",
@@ -239,11 +137,9 @@
}
],
"source": [
- "# 0 no ablation, 1 ablation L2, 2 ablation init, 3 both\n",
- "for abla in range(4):\n",
- " for i in range(5):\n",
- " cmd = 'python ./binary_model/main.py --dataTrain '+ path+'data/cdbpr_format/'+datasets[i]+'/train.csv --dataTest '+path+'data/'+datasets[i]+'/test.csv --ablation '+str(abla)\n",
- " os.system(cmd)"
+ "print(os.getcwd())\n",
+ "cmd = 'cd code/binary_bpr_ablation && python script_ablation.py '\n",
+ "!{cmd}"
]
},
{
@@ -282,9 +178,9 @@
],
"source": [
"# pour all\n",
- "cmd = 'python ./nary_model/main_nary_cv.py --data '+ path+'data/covid/initsurvey.csv' \n",
+ "cmd = 'python code/nary_model/main_nary_cv.py --data '+ path+'data/covid/initsurvey.csv'\n",
"print(os.system(cmd))\n",
- "cmd = 'python ./nary_model/main_nary_cv.py --data '+path+'data/covid/psysurvey.csv' \n",
+ "cmd = 'python code/nary_model/main_nary_cv.py --data '+path+'data/covid/psysurvey.csv'\n",
"print(os.system(cmd))"
]
},
@@ -490,9 +386,9 @@
}
],
"source": [
- "cmd = 'python ./other/decision_tree.py --lower 0'\n",
+ "cmd = 'python code/unsupervised_DT/decision_tree.py --lower 0'\n",
"print(os.system(cmd))\n",
- "cmd = 'python ./other/decision_tree.py --lower 1'\n",
+ "cmd = 'python code/unsupervised_DT/decision_tree.py --lower 1'\n",
"print(os.system(cmd))"
]
},
@@ -552,8 +448,8 @@
"metadata": {},
"outputs": [],
"source": [
- "H_users = fromDFtoArray(path+ \"other/dt/train_embed.csv\",False,'f') \n",
- "classPsy = fromDFtoArray(path+ \"other/dt/train_user_quest_label.csv\",False,'f') "
+ "H_users = fromDFtoArray(path+ \"code/unsupervised_DT/files_for_dt/train_embed.csv\",False,'f')\n",
+ "classPsy = fromDFtoArray(path+ \"code/unsupervised_DT/files_for_dt/train_user_quest_label.csv\",False,'f')"
]
},
{
diff --git a/code/binary_bpr/main.py b/code/binary_bpr/main.py
index 88c9d7d2cc4cc6071350c0c5f5e58f4bdb6b99a1..c9179a8adb1ab1e8c105fcb85dde2d41e63a3a36 100644
--- a/code/binary_bpr/main.py
+++ b/code/binary_bpr/main.py
@@ -62,7 +62,7 @@ def read_file(dataTrain, dataTest):
dico_items = { k:v for (k,v) in zip(items, range(num_items))}
return dico_kc, dico_users, dico_items
- def save_embeddings(xpName: str, modelName: str, embeddings,userEmbDir : str,itemEmbDir : str):
+def save_embeddings(xpName: str, modelName: str, embeddings,userEmbDir : str,itemEmbDir : str):
"""
Saves all the metrics measured after the training process.
diff --git a/code/binary_bpr_ablation/script_abalation.py b/code/binary_bpr_ablation/script_abalation.py
deleted file mode 100644
index 5e56acab0d89733e56e7bde3eede2869f04dd323..0000000000000000000000000000000000000000
--- a/code/binary_bpr_ablation/script_abalation.py
+++ /dev/null
@@ -1,9 +0,0 @@
-import os
-
-name = ["assist09","assist17","algebra", "math1","math2"]
-for i in range(4):
- print("Ablation (0 no ablation, 1 ablation L2, 2 ablation init, 3 both) ",i)
- for a in range(5):
- print(name[i])
- cmd = "python main.py --dataTrain ../data/"+name[i]+"/train.csv --dataTest ../data/"+name[i]+"/test.csv --ablation "+str(i)
- os.system(cmd)
diff --git a/code/binary_bpr_ablation/script_ablation.py b/code/binary_bpr_ablation/script_ablation.py
new file mode 100644
index 0000000000000000000000000000000000000000..9c957a2630526f18c8ab6bfaf368262abd549d21
--- /dev/null
+++ b/code/binary_bpr_ablation/script_ablation.py
@@ -0,0 +1,9 @@
+import os
+
+name = ["assist0910_tkde","assist17_tkde","algebra", "math_1","math_2"]
+for i in range(4):
+ print("Ablation (0 no ablation, 1 ablation L2, 2 ablation init, 3 both) ",i)
+ for a in range(5):
+ print(name[i])
+ cmd = "python main.py --dataTrain ../../data/"+name[i]+"/train_0.csv --dataTest ../../data/"+name[i]+"/test_0.csv --ablation "+str(i)
+ os.system(cmd)
diff --git a/code/nary_model/main_nary_cv.py b/code/nary_model/main_nary_cv.py
index 609ed98d1df6a2b1a7237e1b13951ca1d4415b4d..6816f1d421833630274ad831bffc37d276dc3eac 100644
--- a/code/nary_model/main_nary_cv.py
+++ b/code/nary_model/main_nary_cv.py
@@ -48,7 +48,7 @@ def evaluate_all(dataTrain, dataTest, filename):
new_embedding_items = bpr_model.item_embeddings.weight.clone().detach().cpu().numpy()
write_file(filename+"embedding_items.csv", new_embedding_items[0:nb_item_train])
# Test
- acc, precision, rappel, all_decisions = bpr_model.evaluate_model(test, len(dico_kc), y_test)
+ acc, precision, rappel, all_decisions, all_prefs = bpr_model.evaluate_model(test, len(dico_kc), y_test)
'''
s = str(acc) +","+ str( precision)+ ","+str(rappel)+ ","+str(doa)
for i in range(embedding_size):
diff --git a/code/unsupervised_DT/decision_tree.py b/code/unsupervised_DT/decision_tree.py
index bfba35be1143836458dd6cf1a684add864da03cd..a2f1a73bd8b693b76627a5608f64d45d9699d20f 100644
--- a/code/unsupervised_DT/decision_tree.py
+++ b/code/unsupervised_DT/decision_tree.py
@@ -302,8 +302,8 @@ if __name__ == '__main__':
min_users = 50
os.getcwd()
path = os.path.dirname(os.path.realpath(__file__))
-print(path+ "/dt/train_embed.csv")
-H_users = fromDFtoArray(path+ "/dt/train_embed.csv",False,'f')
+print(path+ "/files_for_dt/train_embed.csv")
+H_users = fromDFtoArray(path+ "/files_for_dt/train_embed.csv",False,'f')
dim = H_users.shape[1]
n = H_users.shape[0]
@@ -327,7 +327,7 @@ for i in range(len(partitions)):
writer.writerow(row)
f.close()
'''
-partitions_init = fromDFtoArray(path+ "/dt/train_partitions.csv",False,'f')
+partitions_init = fromDFtoArray(path+ "/files_for_dt/train_partitions.csv",False,'f')
# Step 2: take the partition with max silhouette value
best_init_part = np.argmax(partitions_init[:,2])
@@ -373,7 +373,7 @@ for i in range(1,len(theClusters)):
# Validation with questionnaires
print("")
-file = path+ "/dt/train_user_quest_label.csv"
+file = path+ "/files_for_dt/train_user_quest_label.csv"
quest = fromDFtoArray(file,False,'f')
questClusters = []
for i in range(len(quest)):
@@ -392,4 +392,4 @@ compare_clusters(dtClust, questClusters, quest)
#print(partitions_init)
-
\ No newline at end of file
+
diff --git a/data.zip b/data.zip
deleted file mode 100644
index 95c40307b2783632095447bfec4bf902cce34e21..0000000000000000000000000000000000000000
Binary files a/data.zip and /dev/null differ