From a28fa020d6c9db0c8cd4c559f9bd859545aadc80 Mon Sep 17 00:00:00 2001
From: sone <sone47.china@gmail.com>
Date: Wed, 25 Aug 2021 20:46:31 +0800
Subject: [PATCH] [fix] answer time processing problem

---
 examples/LPKT/LPKT.ipynb            | 30 ++++++++++++++---------------
 examples/LPKT/LPKT.py               |  4 ++--
 examples/LPKT/prepare_dataset.ipynb | 16 ++++++++-------
 3 files changed, 26 insertions(+), 24 deletions(-)

diff --git a/examples/LPKT/LPKT.ipynb b/examples/LPKT/LPKT.ipynb
index 20b1e7e..b3ba861 100644
--- a/examples/LPKT/LPKT.ipynb
+++ b/examples/LPKT/LPKT.ipynb
@@ -49,8 +49,8 @@
     "    return q_matrix\n",
     "\n",
     "batch_size = 32\n",
-    "n_at = 9632\n",
-    "n_it = 2890\n",
+    "n_at = 1326\n",
+    "n_it = 2839\n",
     "n_question = 102\n",
     "n_exercise = 3162\n",
     "seqlen = 500\n",
@@ -111,52 +111,52 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "Training: 100%|██████████| 31/31 [16:34<00:00, 32.09s/it]\n",
-      "Testing:   0%|          | 0/13 [00:00<?, ?it/s]"
+      "Training: 100%|██████████| 61/61 [17:57<00:00, 17.66s/it]\n",
+      "Testing:   0%|          | 0/26 [00:00<?, ?it/s]"
      ]
     },
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "[Epoch 0] LogisticLoss: 0.747071\n"
+      "[Epoch 0] LogisticLoss: 0.643005\n"
      ]
     },
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "Testing: 100%|██████████| 13/13 [01:14<00:00,  5.76s/it]\n",
-      "Training:   0%|          | 0/31 [00:00<?, ?it/s]"
+      "Testing: 100%|██████████| 26/26 [01:26<00:00,  3.33s/it]\n",
+      "Training:   0%|          | 0/61 [00:00<?, ?it/s]"
      ]
     },
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "[Epoch 0] auc: 0.647351, accuracy: 0.624059\n"
+      "[Epoch 0] auc: 0.680518, accuracy: 0.662649\n"
      ]
     },
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "Training: 100%|██████████| 31/31 [15:22<00:00, 29.76s/it]\n",
-      "Testing:   0%|          | 0/13 [00:00<?, ?it/s]"
+      "Training: 100%|██████████| 61/61 [16:59<00:00, 16.72s/it]\n",
+      "Testing:   0%|          | 0/26 [00:00<?, ?it/s]"
      ]
     },
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "[Epoch 1] LogisticLoss: 0.628299\n"
+      "[Epoch 1] LogisticLoss: 0.596969\n"
      ]
     },
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "Testing: 100%|██████████| 13/13 [01:20<00:00,  6.20s/it]\n",
+      "Testing: 100%|██████████| 26/26 [01:25<00:00,  3.27s/it]\n",
       "INFO:root:save parameters to lpkt.params\n"
      ]
     },
@@ -164,7 +164,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "[Epoch 1] auc: 0.663860, accuracy: 0.670790\n"
+      "[Epoch 1] auc: 0.717534, accuracy: 0.682302\n"
      ]
     }
    ],
@@ -203,14 +203,14 @@
      "output_type": "stream",
      "text": [
       "INFO:root:load parameters from lpkt.params\n",
-      "Testing: 100%|██████████| 13/13 [01:20<00:00,  6.23s/it]"
+      "Testing: 100%|██████████| 26/26 [01:18<00:00,  3.04s/it]"
      ]
     },
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "auc: 0.670790, accuracy: 0.663860\n"
+      "auc: 0.717543, accuracy: 0.682427\n"
      ]
     },
     {
diff --git a/examples/LPKT/LPKT.py b/examples/LPKT/LPKT.py
index 591353c..6bdccd3 100644
--- a/examples/LPKT/LPKT.py
+++ b/examples/LPKT/LPKT.py
@@ -13,8 +13,8 @@ def generate_q_matrix(path, n_skill, n_problem, gamma=0.0):
     return q_matrix
 
 batch_size = 32
-n_at = 9632
-n_it = 2890
+n_at = 1326
+n_it = 2839
 n_question = 102
 n_exercise = 3162
 seqlen = 500
diff --git a/examples/LPKT/prepare_dataset.ipynb b/examples/LPKT/prepare_dataset.ipynb
index 541e145..b9c1280 100644
--- a/examples/LPKT/prepare_dataset.ipynb
+++ b/examples/LPKT/prepare_dataset.ipynb
@@ -30,7 +30,9 @@
     "data = pd.read_csv(\n",
     "    '../../data/anonymized_full_release_competition_dataset/anonymized_full_release_competition_dataset.csv',\n",
     "    usecols=['startTime', 'timeTaken', 'studentId', 'skill', 'problemId', 'correct']\n",
-    ").dropna(subset=['skill', 'problemId']).sort_values('startTime')"
+    ").dropna(subset=['skill', 'problemId']).sort_values('startTime')\n",
+    "\n",
+    "data.timeTaken = data.timeTaken.astype(int)"
    ]
   },
   {
@@ -44,7 +46,7 @@
      "text": [
       "number of skills: 102\n",
       "number of problems: 3162\n",
-      "number of answer time: 9632\n"
+      "number of answer time: 1326\n"
      ]
     }
    ],
@@ -56,7 +58,7 @@
     "# question id from 1 to #num_skill\n",
     "skill2id = { p: i+1 for i, p in enumerate(skills) }\n",
     "problem2id = { p: i+1 for i, p in enumerate(problems) }\n",
-    "at2id = { int(a): i + 1 for i, a in enumerate(at) }\n",
+    "at2id = { a: i for i, a in enumerate(at) }\n",
     "\n",
     "print(\"number of skills: %d\" % len(skills))\n",
     "print(\"number of problems: %d\" % len(problems))\n",
@@ -88,7 +90,7 @@
     "            item = 43200\n",
     "        it.add(item)\n",
     "\n",
-    "it2id = { a: i + 1 for i, a in enumerate(it) }\n",
+    "it2id = { a: i for i, a in enumerate(it) }\n",
     "print(\"number of interval time: %d\" % len(it))"
    ],
    "metadata": {
@@ -126,7 +128,7 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "parse student sequence:\t: 100%|██████████| 1709/1709 [00:03<00:00, 495.23it/s]\n"
+      "parse student sequence:\t: 100%|██████████| 1709/1709 [00:03<00:00, 534.71it/s]\n"
      ]
     }
    ],
@@ -187,8 +189,8 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "write into file: 100%|██████████| 1196/1196 [00:00<00:00, 1638.25it/s]\n",
-      "write into file: 100%|██████████| 513/513 [00:00<00:00, 1795.56it/s]\n"
+      "write into file: 100%|██████████| 1196/1196 [00:00<00:00, 1661.06it/s]\n",
+      "write into file: 100%|██████████| 513/513 [00:00<00:00, 1757.38it/s]\n"
      ]
     }
    ],