From 6140d8d228b1d428c1b13baca6581ffbd6a04e8a Mon Sep 17 00:00:00 2001
From: Filip Johnsson <filjo653@student.liu.se>
Date: Mon, 25 Nov 2024 16:19:47 +0100
Subject: [PATCH] lab3 working
---
l3/TM-Lab3.ipynb | 222 +++++++++++++++++++++++++++++++++++++++++++----
1 file changed, 205 insertions(+), 17 deletions(-)
diff --git a/l3/TM-Lab3.ipynb b/l3/TM-Lab3.ipynb
index b14e058..1e0074f 100644
--- a/l3/TM-Lab3.ipynb
+++ b/l3/TM-Lab3.ipynb
@@ -16727,7 +16727,7 @@
},
{
"cell_type": "code",
- "execution_count": 58,
+ "execution_count": 62,
"metadata": {
"deletable": false,
"nbgrader": {
@@ -16784,7 +16784,7 @@
},
{
"cell_type": "code",
- "execution_count": 61,
+ "execution_count": 63,
"metadata": {
"deletable": false,
"editable": false,
@@ -16803,7 +16803,27 @@
"solution"
]
},
- "outputs": [],
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Precision: 0.668, Recall: 0.610, F1: 0.638\n"
+ ]
+ },
+ {
+ "data": {
+ "text/html": [
+ "<div class=\"alert alert-success\"><strong>Checks have passed!</strong></div>"
+ ],
+ "text/plain": [
+ "<IPython.core.display.HTML object>"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
"source": [
"scores = evaluation_scores(dev_gold_mentions, set(most_probable_method(df_dev_pred, df_kb)))\n",
"print_evaluation_scores(scores)\n",
@@ -16831,7 +16851,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 64,
"metadata": {
"deletable": false,
"editable": false,
@@ -16861,9 +16881,91 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 65,
"metadata": {},
- "outputs": [],
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "<div>\n",
+ "<style scoped>\n",
+ " .dataframe tbody tr th:only-of-type {\n",
+ " vertical-align: middle;\n",
+ " }\n",
+ "\n",
+ " .dataframe tbody tr th {\n",
+ " vertical-align: top;\n",
+ " }\n",
+ "\n",
+ " .dataframe thead th {\n",
+ " text-align: right;\n",
+ " }\n",
+ "</style>\n",
+ "<table border=\"1\" class=\"dataframe\">\n",
+ " <thead>\n",
+ " <tr style=\"text-align: right;\">\n",
+ " <th></th>\n",
+ " <th>mention</th>\n",
+ " <th>entity</th>\n",
+ " <th>context</th>\n",
+ " </tr>\n",
+ " </thead>\n",
+ " <tbody>\n",
+ " <tr>\n",
+ " <th>0</th>\n",
+ " <td>1970</td>\n",
+ " <td>UEFA_Champions_League</td>\n",
+ " <td>Cup twice the first in @ and the second in 1983</td>\n",
+ " </tr>\n",
+ " <tr>\n",
+ " <th>1</th>\n",
+ " <td>1970</td>\n",
+ " <td>FIFA_World_Cup</td>\n",
+ " <td>America 1975 and during the @ and 1978 World C...</td>\n",
+ " </tr>\n",
+ " <tr>\n",
+ " <th>2</th>\n",
+ " <td>1990 World Cup</td>\n",
+ " <td>1990_FIFA_World_Cup</td>\n",
+ " <td>Manolo represented Spain at the @</td>\n",
+ " </tr>\n",
+ " <tr>\n",
+ " <th>3</th>\n",
+ " <td>1990 World Cup</td>\n",
+ " <td>1990_FIFA_World_Cup</td>\n",
+ " <td>Hašek represented Czechoslovakia at the @ and ...</td>\n",
+ " </tr>\n",
+ " <tr>\n",
+ " <th>4</th>\n",
+ " <td>1990 World Cup</td>\n",
+ " <td>1990_FIFA_World_Cup</td>\n",
+ " <td>renovations in 1989 for the @ The present capa...</td>\n",
+ " </tr>\n",
+ " </tbody>\n",
+ "</table>\n",
+ "</div>"
+ ],
+ "text/plain": [
+ " mention entity \\\n",
+ "0 1970 UEFA_Champions_League \n",
+ "1 1970 FIFA_World_Cup \n",
+ "2 1990 World Cup 1990_FIFA_World_Cup \n",
+ "3 1990 World Cup 1990_FIFA_World_Cup \n",
+ "4 1990 World Cup 1990_FIFA_World_Cup \n",
+ "\n",
+ " context \n",
+ "0 Cup twice the first in @ and the second in 1983 \n",
+ "1 America 1975 and during the @ and 1978 World C... \n",
+ "2 Manolo represented Spain at the @ \n",
+ "3 Hašek represented Czechoslovakia at the @ and ... \n",
+ "4 renovations in 1989 for the @ The present capa... "
+ ]
+ },
+ "execution_count": 65,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
"source": [
"df_contexts.head()"
]
@@ -16879,9 +16981,31 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 66,
"metadata": {},
- "outputs": [],
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "41465 Nebraska Concealed Handgun Permit In @ municip...\n",
+ "41466 Lazlo restaurants are located in @ and Omaha C...\n",
+ "41467 California Washington Overland Park Kansas @ N...\n",
+ "41468 City Missouri Omaha Nebraska and @ Nebraska It...\n",
+ "41469 by Sandhills Publishing Company in @ Nebraska USA\n",
+ " ... \n",
+ "41609 @ Leyton Orient\n",
+ "41610 English division three Swansea @ \n",
+ "41611 league membership narrowly edging out @ on goa...\n",
+ "41612 @ Cambridge\n",
+ "41613 @ \n",
+ "Name: context, Length: 149, dtype: object"
+ ]
+ },
+ "execution_count": 66,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
"source": [
"df_contexts.context[df_contexts.mention == 'Lincoln']"
]
@@ -16901,7 +17025,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 105,
"metadata": {
"deletable": false,
"nbgrader": {
@@ -16920,6 +17044,10 @@
},
"outputs": [],
"source": [
+ "from sklearn.naive_bayes import MultinomialNB\n",
+ "from sklearn.pipeline import Pipeline\n",
+ "from sklearn.feature_extraction.text import CountVectorizer\n",
+ "\n",
"def build_entity_classifiers(df_kb, df_contexts):\n",
" \"\"\"Build Naive Bayes classifiers for entity prediction.\n",
"\n",
@@ -16932,13 +17060,45 @@
" classifiers trained to predict the correct entity, given the textual\n",
" context of the mention (as described in detail above).\n",
" \"\"\"\n",
- " # YOUR CODE HERE\n",
- " raise NotImplementedError()"
+ " # === Get X and Y training data ===\n",
+ "\n",
+ " mention_training_data_dict = {}\n",
+ "\n",
+ " for row in df_contexts.itertuples():\n",
+ " mention = row[1]\n",
+ " entity = row[2]\n",
+ " context = row[3]\n",
+ "\n",
+ " if mention not in mention_training_data_dict:\n",
+ " mention_training_data_dict[mention] = {'trainX': [], 'trainY': []}\n",
+ " \n",
+ " mention_training_data_dict[mention]['trainX'].append(context)\n",
+ " mention_training_data_dict[mention]['trainY'].append(entity)\n",
+ " \n",
+ " # === Create pipelines ===\n",
+ "\n",
+ " mention_classifier_dict = {}\n",
+ "\n",
+ " for mention, training_data in mention_training_data_dict.items():\n",
+ " probs = df_kb.loc[df_kb.mention == mention].sort_values(\"entity\")\n",
+ " class_prior = list(probs[\"prob\"])\n",
+ "\n",
+ " pipeline = Pipeline(steps=[\n",
+ " ('vectorizer', CountVectorizer()),\n",
+ " ('classifier', MultinomialNB(class_prior=class_prior)),\n",
+ " ])\n",
+ " trainX = np.array(training_data['trainX'])\n",
+ " trainY = np.array(training_data['trainY'])\n",
+ " model = pipeline.fit(trainX, trainY)\n",
+ "\n",
+ " mention_classifier_dict[mention] = model\n",
+ " \n",
+ " return mention_classifier_dict"
]
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 106,
"metadata": {
"deletable": false,
"nbgrader": {
@@ -16972,8 +17132,28 @@
" quadruples consisting of the sentence id, start position, end\n",
" position and the predicted entity label of each span.\n",
" \"\"\"\n",
- " # YOUR CODE HERE\n",
- " raise NotImplementedError()"
+ " for row in df.itertuples():\n",
+ " sentence_id = row[1]\n",
+ " sentence = row[2]\n",
+ " start = row[3]\n",
+ " end = row[4]\n",
+ "\n",
+ " words = sentence.split()\n",
+ "\n",
+ " mention = ' '.join(words[int(start):int(end)])\n",
+ " context = ' '.join(words[:int(start)]+['@']+words[int(end):])\n",
+ " \n",
+ " if mention in classifiers:\n",
+ " classifier = classifiers[mention]\n",
+ " predicted_entity = classifier.predict([context])[0]\n",
+ " \n",
+ " else:\n",
+ " entity = df_kb.loc[df_kb.mention == mention].entity\n",
+ " predicted_entity = entity.values[0] if len(entity.values) > 0 else \"--NME--\"\n",
+ "\n",
+ " yield sentence_id, start, end, predicted_entity\n",
+ "\n",
+ " "
]
},
{
@@ -16987,7 +17167,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 107,
"metadata": {
"deletable": false,
"editable": false,
@@ -17018,7 +17198,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 108,
"metadata": {
"deletable": false,
"editable": false,
@@ -17037,7 +17217,15 @@
"solution"
]
},
- "outputs": [],
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Precision: 0.686, Recall: 0.627, F1: 0.655\n"
+ ]
+ }
+ ],
"source": [
"scores = evaluation_scores(dev_gold_mentions, dev_pred_dict_mentions)\n",
"print_evaluation_scores(scores)"
--
GitLab