{ "cells": [ { "cell_type": "markdown", "id": "f40acb6f", "metadata": {}, "source": [ "# Time series classification for dropout prediction\n", "\n", "We adopt the time series classification methodology presented in the work of Liu et\n", "al. {cite}`liu_2018` for student dropout prediction.\n", "\n", "Time-series classifiers can effectively visualize and categorize data based on\n", "pattern similarities, enabling them to learn distinctive series of behaviors,\n", "distinguishing dropout from retention.\n", "\n", "The classification approach uses student interaction data (click-stream) on\n", "different resource types ('resource', 'content' and 'forumng').\n", "Where 'oucontent' and 'resource' refer to a lecture video and a segment of text\n", "students are supposed to watch or read, and 'forumng' points to the forum space of\n", "the course.\n", "\n", "A student is considered a dropout if he withdraws from a course.\n", "In the OULAD dataset, this information is recorded with a `Withdrawn` value in the\n", "`final_result` column of the `student_info` table.\n", "\n", "The authors rearranged the OULAD interaction data and transformed it into several\n", "time series using the following steps:\n", "\n", "1. Extract the number of clicks the students make on the three types of material and\n", " group them by course module presentation.\n", "2. Sum up the number of clicks each student makes on each type of material from each\n", " presentation daily.\n", "3. Align each student’s total clicks on each type of material by days.\n", "4. Add the dropout label, withdrawn as `1`, otherwise as `0`, to the end of each\n", " student instance.\n", "\n", "Thus, three time-series datasets are constructed for each course module presentation.\n", "\n", "Finally, student dropout is predicted by applying a Time series forest classifier on\n", "each time series by course.\n", "\n", "```{bibliography}\n", ":filter: docname in docnames\n", "```" ] }, { "cell_type": "code", "execution_count": 1, "id": "1a4049aa", "metadata": {}, "outputs": [], "source": [ "from typing import Iterator\n", "\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import pandas as pd\n", "from IPython.display import Markdown, display\n", "from sklearn.model_selection import cross_val_score\n", "from sktime.classification.interval_based import TimeSeriesForestClassifier\n", "\n", "from oulad import get_oulad\n", "\n", "%load_ext oulad.capture" ] }, { "cell_type": "code", "execution_count": 2, "id": "9fd68797", "metadata": {}, "outputs": [], "source": [ "%%capture oulad\n", "oulad = get_oulad()" ] }, { "cell_type": "markdown", "id": "37d4eba4", "metadata": {}, "source": [ "## Dropout rate by course presentation\n", "\n", "As we embark on the task of student dropout classification, it's important to first\n", "familiarize ourselves with some basic descriptive statistics about the dataset.\n", "As demonstrated in the work of Lui et al., a concise summary of domain information\n", "and relevant statistics regarding student dropout in diverse course modules can prove\n", "insightful.\n", "We shall attempt to replicate this table in our present study." ] }, { "cell_type": "code", "execution_count": 3, "id": "7cf00e6e", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
code_modulecode_presentationdomainstudentsdaysdropout
0AAA2013JSocial Sciences38327915.67%
1AAA2014JSocial Sciences36529418.08%
2BBB2013BSocial Sciences176725028.58%
3BBB2013JSocial Sciences223729228.79%
4BBB2014BSocial Sciences161324430.38%
5BBB2014JSocial Sciences229227232.68%
6CCC2014BSTEM193626046.38%
7CCC2014JSTEM249828843.11%
8DDD2013BSTEM130325733.15%
9DDD2013JSTEM193828035.14%
10DDD2014BSTEM122826039.90%
11DDD2014JSTEM180328835.88%
12EEE2013JSTEM105228023.10%
13EEE2014BSTEM69426024.93%
14EEE2014JSTEM118828825.76%
15FFF2013BSTEM161425925.46%
16FFF2013JSTEM228328729.57%
17FFF2014BSTEM150026030.80%
18FFF2014JSTEM236528836.15%
19GGG2013JSocial Sciences9522786.93%
20GGG2014BSocial Sciences83325012.00%
21GGG2014JSocial Sciences74928616.82%
\n", "
" ], "text/plain": [ " code_module code_presentation domain students days dropout\n", "0 AAA 2013J Social Sciences 383 279 15.67%\n", "1 AAA 2014J Social Sciences 365 294 18.08%\n", "2 BBB 2013B Social Sciences 1767 250 28.58%\n", "3 BBB 2013J Social Sciences 2237 292 28.79%\n", "4 BBB 2014B Social Sciences 1613 244 30.38%\n", "5 BBB 2014J Social Sciences 2292 272 32.68%\n", "6 CCC 2014B STEM 1936 260 46.38%\n", "7 CCC 2014J STEM 2498 288 43.11%\n", "8 DDD 2013B STEM 1303 257 33.15%\n", "9 DDD 2013J STEM 1938 280 35.14%\n", "10 DDD 2014B STEM 1228 260 39.90%\n", "11 DDD 2014J STEM 1803 288 35.88%\n", "12 EEE 2013J STEM 1052 280 23.10%\n", "13 EEE 2014B STEM 694 260 24.93%\n", "14 EEE 2014J STEM 1188 288 25.76%\n", "15 FFF 2013B STEM 1614 259 25.46%\n", "16 FFF 2013J STEM 2283 287 29.57%\n", "17 FFF 2014B STEM 1500 260 30.80%\n", "18 FFF 2014J STEM 2365 288 36.15%\n", "19 GGG 2013J Social Sciences 952 278 6.93%\n", "20 GGG 2014B Social Sciences 833 250 12.00%\n", "21 GGG 2014J Social Sciences 749 286 16.82%" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "display(\n", " # Add code_module, code_presentation, module_presentation_length columns.\n", " oulad.courses\n", " # Add domain column.\n", " .merge(oulad.domains, on=\"code_module\")\n", " # Add days column.\n", " .merge(\n", " oulad.student_vle[oulad.student_vle.date < 0]\n", " .groupby([\"code_module\", \"code_presentation\"])[\"date\"]\n", " .min()\n", " .reset_index(),\n", " on=[\"code_module\", \"code_presentation\"],\n", " )\n", " .assign(days=lambda df: df.module_presentation_length - df.date + 1)\n", " # Add id_student and final_result columns.\n", " .merge(\n", " oulad.student_info[\n", " [\"id_student\", \"code_module\", \"code_presentation\", \"final_result\"]\n", " ],\n", " on=[\"code_module\", \"code_presentation\"],\n", " )\n", " .assign(final_result=lambda df: df.final_result == \"Withdrawn\")\n", " # Aggregate by course, rename final_result to dropout and id_student to students.\n", " .groupby([\"code_module\", \"code_presentation\"])\n", " .agg(\n", " domain=(\"domain\", \"first\"),\n", " students=(\"id_student\", \"nunique\"),\n", " days=(\"days\", \"first\"),\n", " dropout=(\"final_result\", \"mean\"),\n", " )\n", " # Format the dropout column using percentages.\n", " .assign(dropout=lambda df: df.dropout.apply(lambda x: f\"{x * 100:.2f}%\"))\n", " .reset_index()\n", ")" ] }, { "cell_type": "markdown", "id": "b71ee65a", "metadata": {}, "source": [ "## Prepare interaction data\n", "\n", "In this section, we attempt to replicate the interaction data transformations\n", "as described in the work of Liu et al." ] }, { "cell_type": "code", "execution_count": 4, "id": "3be01f07", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
-25-24-23-22-21-20-19-18-17-16...261262263264265266267268269final_result
code_modulecode_presentationid_studentactivity_type
AAA2013J11391forumng0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.00
oucontent0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.00
resource0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.00
28400forumng0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.00
oucontent0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.00
...........................................................................
GGG2014J2679821oucontent0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.01
resource0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.01
2684003forumng0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.00
oucontent0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.00
resource0.00.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.00
\n", "

80524 rows × 296 columns

\n", "
" ], "text/plain": [ " -25 -24 -23 -22 \\\n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 11391 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", " 28400 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", "... ... ... ... ... \n", "GGG 2014J 2679821 oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", " 2684003 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", "\n", " -21 -20 -19 -18 \\\n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 11391 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", " 28400 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", "... ... ... ... ... \n", "GGG 2014J 2679821 oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", " 2684003 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", "\n", " -17 -16 ... 261 \\\n", "code_module code_presentation id_student activity_type ... \n", "AAA 2013J 11391 forumng 0.0 0.0 ... 0.0 \n", " oucontent 0.0 0.0 ... 0.0 \n", " resource 0.0 0.0 ... 0.0 \n", " 28400 forumng 0.0 0.0 ... 0.0 \n", " oucontent 0.0 0.0 ... 0.0 \n", "... ... ... ... ... \n", "GGG 2014J 2679821 oucontent 0.0 0.0 ... 0.0 \n", " resource 0.0 0.0 ... 0.0 \n", " 2684003 forumng 0.0 0.0 ... 0.0 \n", " oucontent 0.0 0.0 ... 0.0 \n", " resource 0.0 0.0 ... 0.0 \n", "\n", " 262 263 264 265 \\\n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 11391 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", " 28400 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", "... ... ... ... ... \n", "GGG 2014J 2679821 oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", " 2684003 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", "\n", " 266 267 268 269 \\\n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 11391 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", " 28400 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", "... ... ... ... ... \n", "GGG 2014J 2679821 oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", " 2684003 forumng 0.0 0.0 0.0 0.0 \n", " oucontent 0.0 0.0 0.0 0.0 \n", " resource 0.0 0.0 0.0 0.0 \n", "\n", " final_result \n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 11391 forumng 0 \n", " oucontent 0 \n", " resource 0 \n", " 28400 forumng 0 \n", " oucontent 0 \n", "... ... \n", "GGG 2014J 2679821 oucontent 1 \n", " resource 1 \n", " 2684003 forumng 0 \n", " oucontent 0 \n", " resource 0 \n", "\n", "[80524 rows x 296 columns]" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "%%capture -ns time_series_classification_for_dropout_prediction click_stream\n", "click_stream = (\n", " # 1. Extract the number of clicks by students on the three types of material.\n", " oulad.vle.query(\"activity_type in ['resource', 'oucontent', 'forumng']\")\n", " .drop([\"code_module\", \"code_presentation\", \"week_from\", \"week_to\"], axis=1)\n", " .merge(oulad.student_vle, on=\"id_site\")\n", " # 2. Sum the number of clicks each student makes on each type of material by day.\n", " .groupby(\n", " [\"code_module\", \"code_presentation\", \"id_student\", \"activity_type\", \"date\"]\n", " )\n", " .agg({\"sum_click\": \"sum\"})\n", " # 3. Align each student’s total clicks on each type of material by days.\n", " .pivot_table(\n", " values=\"sum_click\",\n", " index=[\"code_module\", \"code_presentation\", \"id_student\", \"activity_type\"],\n", " columns=\"date\",\n", " fill_value=0.0,\n", " )\n", " # 4. Add the dropout label, withdrawn as `1`, otherwise as `0`.\n", " .join(\n", " oulad.student_info.filter(\n", " [\"code_module\", \"code_presentation\", \"id_student\", \"final_result\"]\n", " )\n", " .assign(final_result=lambda df: (df.final_result == \"Withdrawn\").astype(int))\n", " .set_index([\"code_module\", \"code_presentation\", \"id_student\"])\n", " )\n", ")\n", "display(click_stream)" ] }, { "cell_type": "markdown", "id": "06fcabba", "metadata": {}, "source": [ "### Example\n", "\n", "Presented below is an extract of student interactions conducted within the\n", "\"AAA_2013J\" course presentation, specifically pertaining to forum activities.\n", "This extract corresponds to Table 2 as illustrated in the work of Lui et al." ] }, { "cell_type": "code", "execution_count": 5, "id": "8d2af30f", "metadata": { "lines_to_next_cell": 2 }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
-10-9-8-7-6-5-4-3-2-1...261262263264265266267268269final_result
code_modulecode_presentationid_studentactivity_type
AAA2013J28400forumng14.00.04.017.08.03.04.00.023.00.0...0.00.00.00.00.00.00.00.00.00
30268forumng5.00.02.00.00.00.00.09.017.06.0...0.00.00.00.00.00.00.00.00.01
31604forumng0.06.00.00.018.00.00.00.05.00.0...0.00.00.00.00.00.00.00.00.00
2694424forumng0.01.00.00.00.00.00.00.00.00.0...0.00.00.00.00.00.00.00.00.00
\n", "

4 rows × 281 columns

\n", "
" ], "text/plain": [ " -10 -9 -8 -7 \\\n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 28400 forumng 14.0 0.0 4.0 17.0 \n", " 30268 forumng 5.0 0.0 2.0 0.0 \n", " 31604 forumng 0.0 6.0 0.0 0.0 \n", " 2694424 forumng 0.0 1.0 0.0 0.0 \n", "\n", " -6 -5 -4 -3 \\\n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 28400 forumng 8.0 3.0 4.0 0.0 \n", " 30268 forumng 0.0 0.0 0.0 9.0 \n", " 31604 forumng 18.0 0.0 0.0 0.0 \n", " 2694424 forumng 0.0 0.0 0.0 0.0 \n", "\n", " -2 -1 ... 261 \\\n", "code_module code_presentation id_student activity_type ... \n", "AAA 2013J 28400 forumng 23.0 0.0 ... 0.0 \n", " 30268 forumng 17.0 6.0 ... 0.0 \n", " 31604 forumng 5.0 0.0 ... 0.0 \n", " 2694424 forumng 0.0 0.0 ... 0.0 \n", "\n", " 262 263 264 265 \\\n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 28400 forumng 0.0 0.0 0.0 0.0 \n", " 30268 forumng 0.0 0.0 0.0 0.0 \n", " 31604 forumng 0.0 0.0 0.0 0.0 \n", " 2694424 forumng 0.0 0.0 0.0 0.0 \n", "\n", " 266 267 268 269 \\\n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 28400 forumng 0.0 0.0 0.0 0.0 \n", " 30268 forumng 0.0 0.0 0.0 0.0 \n", " 31604 forumng 0.0 0.0 0.0 0.0 \n", " 2694424 forumng 0.0 0.0 0.0 0.0 \n", "\n", " final_result \n", "code_module code_presentation id_student activity_type \n", "AAA 2013J 28400 forumng 0 \n", " 30268 forumng 1 \n", " 31604 forumng 0 \n", " 2694424 forumng 0 \n", "\n", "[4 rows x 281 columns]" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "display(\n", " click_stream.loc[\n", " [\n", " (\"AAA\", \"2013J\", 28400, \"forumng\"),\n", " (\"AAA\", \"2013J\", 30268, \"forumng\"),\n", " (\"AAA\", \"2013J\", 31604, \"forumng\"),\n", " (\"AAA\", \"2013J\", 2694424, \"forumng\"),\n", " ],\n", " -10:\"final_result\",\n", " ]\n", ")" ] }, { "cell_type": "markdown", "id": "76dcdd24", "metadata": { "lines_to_next_cell": 2 }, "source": [ "## Dropout prediction by course\n", "\n", "Using the obtained `click_stream` dataset, we train a Time series Forest\n", "Classifier predicting students dropout for each course presentation.\n", "\n", "As in the work of Lui et al. we set the number of trees in the Time series Forest to\n", "500 and perform 10-fold cross-validation for each course module presentation using\n", "the classification accuracy as the evaluation metric." ] }, { "cell_type": "code", "execution_count": 6, "id": "35aa07c9", "metadata": { "lines_to_next_cell": 2 }, "outputs": [ { "data": { "text/markdown": [ "### Dropout prediction 10-fold cross-validation accuracy by course" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
activity_typeforumngoucontentresource
code_modulecode_presentation
AAA2013J0.8928170.9364860.884282
2014J0.8689080.8927780.878235
BBB2013B0.8183260.7794890.808844
2013J0.8502920.8745220.832293
2014B0.8431030.8690540.846553
2014J0.8206900.8432650.830800
CCC2014B0.7678160.7435360.831991
2014J0.8209590.7780660.878427
DDD2013B0.8218520.7861550.776506
2013J0.8378870.8354850.823526
2014B0.7783540.7864640.807477
2014J0.8279890.8714710.865274
EEE2013J0.8584270.8439310.893902
2014B0.8284150.8217870.905333
2014J0.8846260.8644340.918364
FFF2013B0.8358110.8361550.853654
2013J0.8217630.8262140.825844
2014B0.7860100.8166780.810618
2014J0.8527640.8944280.874186
GGG2013J0.9596220.9432640.940909
2014B0.9152540.9100900.917649
2014J0.8784090.8862510.850469
\n", "
" ], "text/plain": [ "activity_type forumng oucontent resource\n", "code_module code_presentation \n", "AAA 2013J 0.892817 0.936486 0.884282\n", " 2014J 0.868908 0.892778 0.878235\n", "BBB 2013B 0.818326 0.779489 0.808844\n", " 2013J 0.850292 0.874522 0.832293\n", " 2014B 0.843103 0.869054 0.846553\n", " 2014J 0.820690 0.843265 0.830800\n", "CCC 2014B 0.767816 0.743536 0.831991\n", " 2014J 0.820959 0.778066 0.878427\n", "DDD 2013B 0.821852 0.786155 0.776506\n", " 2013J 0.837887 0.835485 0.823526\n", " 2014B 0.778354 0.786464 0.807477\n", " 2014J 0.827989 0.871471 0.865274\n", "EEE 2013J 0.858427 0.843931 0.893902\n", " 2014B 0.828415 0.821787 0.905333\n", " 2014J 0.884626 0.864434 0.918364\n", "FFF 2013B 0.835811 0.836155 0.853654\n", " 2013J 0.821763 0.826214 0.825844\n", " 2014B 0.786010 0.816678 0.810618\n", " 2014J 0.852764 0.894428 0.874186\n", "GGG 2013J 0.959622 0.943264 0.940909\n", " 2014B 0.915254 0.910090 0.917649\n", " 2014J 0.878409 0.886251 0.850469" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "%%capture -ns time_series_classification_for_dropout_prediction results\n", "def get_scores_by_activity_type() -> Iterator[list[float]]:\n", " \"\"\"Computes accuracy prediction scores for each course.\"\"\"\n", " for levels, course_df in click_stream.groupby(\n", " level=[\"code_module\", \"code_presentation\", \"activity_type\"]\n", " ):\n", " estimator = TimeSeriesForestClassifier(n_estimators=500)\n", " X = course_df.drop(columns=\"final_result\").values\n", " y = course_df[\"final_result\"].values\n", " mean_score = np.mean(\n", " cross_val_score(estimator, X, y, cv=10, scoring=\"accuracy\", n_jobs=-1)\n", " )\n", " yield list(levels) + [mean_score]\n", "\n", "\n", "results = pd.DataFrame(\n", " list(get_scores_by_activity_type()),\n", " columns=[\"code_module\", \"code_presentation\", \"activity_type\", \"score\"],\n", ").pivot_table(\n", " values=\"score\",\n", " index=[\"code_module\", \"code_presentation\"],\n", " columns=\"activity_type\",\n", ")\n", "display(Markdown(\"### Dropout prediction 10-fold cross-validation accuracy by course\"))\n", "display(results)" ] }, { "cell_type": "markdown", "id": "4b6bf387", "metadata": { "lines_to_next_cell": 2 }, "source": [ "## Early dropout prediction\n", "\n", "The earlier we can accurately forecast student dropout, the more beneficial the\n", "approach becomes by affording MOOC instructors extra time for intervention.\n", "As in the work of Lui et al., we compare the predictive accuracy of the\n", "TimeSeriesForest classification using incrementally larger fractions of the original\n", "data.\n", "We start with a 5 percent dataset and iteratively add 5 percent increments to assess\n", "how prediction accuracy evolves." ] }, { "cell_type": "code", "execution_count": 8, "id": "2b28bf3f", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "%%capture -ns time_series_classification_for_dropout_prediction a_scores c_scores\n", "def get_score_by_slice(index_slice) -> Iterator[float]:\n", " \"\"\"Computes accuracy prediction scores for the given `index_slice`.\"\"\"\n", " y = click_stream.loc[index_slice, \"final_result\"].values\n", " for i in range(5, 100, 5):\n", " estimator = TimeSeriesForestClassifier(n_estimators=500)\n", " limit = round((click_stream.columns.shape[0] - 1) * i / 100)\n", " X = click_stream.loc[index_slice, click_stream.columns[:limit]].values\n", " yield np.mean(\n", " cross_val_score(estimator, X, y, cv=10, scoring=\"accuracy\", n_jobs=-1)\n", " )\n", "\n", "\n", "a_scores = pd.Series(\n", " list(get_score_by_slice((\"AAA\", \"2013J\", slice(None), \"oucontent\"))),\n", " index=range(5, 100, 5),\n", ")\n", "c_scores = pd.Series(\n", " list(get_score_by_slice((\"CCC\", \"2014J\", slice(None), \"resource\"))),\n", " index=range(5, 100, 5),\n", ")\n", "for scores, name in [(a_scores, \"AAA2013Joucontent\"), (c_scores, \"CCC2014Jresource\")]:\n", " scores.plot(\n", " title=f\"Accuracy score by the percentage of data used for {name} course\",\n", " xlabel=\"Percentage of data used\",\n", " ylabel=\"Prediction accuracy\",\n", " xticks=range(5, 100, 5),\n", " figsize=(10, 5),\n", " grid=True,\n", " marker=\"s\",\n", " )\n", " plt.show()" ] } ], "metadata": { "jupytext": { "cell_metadata_filter": "-all", "main_language": "python", "notebook_metadata_filter": "-all" }, "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.12.7" } }, "nbformat": 4, "nbformat_minor": 5 }