jonas/nlp-lab
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
d899e2f069
nlp-lab/Carsten_Solutions/Task 1 - Carsten Draschner.ipynb
Carsten 322200b2e0 update
2018-06-04 16:57:09 +02:00

1322 lines
36 KiB
Plaintext
Raw Blame History

{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Exercise 1\n",
"Solution by:\n",
"Carsten Draschner \n",
"2719095\n",
"\n",
"Following Instructions: \n",
"https://github.com/SmartDataAnalytics/MA-INF-4222-NLP-Lab/blob/master/2018_SoSe/exercises/Task01_Instructions.ipynb"
]
},
{
"cell_type": "code",
"execution_count": 152,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"import numpy as np\n",
"import nltk\n",
"from nltk import word_tokenize, pos_tag"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Classifiers\n",
"**Decision Tree** import from skikit learn"
]
},
{
"cell_type": "code",
"execution_count": 153,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"from sklearn.tree import DecisionTreeClassifier\n",
"from sklearn.feature_extraction import DictVectorizer\n",
"from sklearn.pipeline import Pipeline"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 1. model1 = your POS tagger model (english)\n",
"for a words defined by its in dex with the given sentences a feature vector fot this word will be determinded"
]
},
{
"cell_type": "code",
"execution_count": 154,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"defined own feature model\n"
]
}
],
"source": [
"def features(sentence, index):\n",
" return {\n",
" 'word': sentence[index],\n",
" 'length': len(sentence[index]),\n",
" 'is_capitalized': sentence[index][0].upper() == sentence[index][0],\n",
" 'prefix-1': sentence[index][0],\n",
" 'suffix-1': sentence[index][-1],\n",
" 'prev_word': '' if index == 0 else sentence[index - 1],\n",
" 'next_word': '' if index == len(sentence) - 1 else sentence[index + 1],\n",
" 'kindOfCamelCase': sentence[index][1:].lower() != sentence[index][1:],\n",
" 'includesSpace': True if ((' ') in sentence[index]) else False, #depemds on tokenizer\n",
" 'containsNumber': sum(str(i) in (sentence[index]) for i in range(10))>0,\n",
" 'prefix-2': sentence[index][1] if len(sentence[index])>1 else \"-1\",\n",
" 'suffix-2': sentence[index][-2] if len(sentence[index])>1 else \"-1\"\n",
" }\n",
"print(\"defined own feature model\")\n",
"#print(features(\"halli hallo i bims der Programmierer\".strip().split(\" \"), 3))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 2. model2 = pre-trained POS tagger model using NLTK (maxentropy english)\n"
]
},
{
"cell_type": "code",
"execution_count": 155,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"#max entropie pre trained pos tag\n",
"#see Calculate performance 1.2"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 3. model3.x = rule-based classifiers (x = 1 to 5)\n",
"1. DefaultTagger that simply tags everything with the same tag\n",
"2. RegexpTagger that applies tags according to a set of regular expressions\n",
"3. N-Gram (n-gram tagger is a generalization of a unigram tagger whose context is the current word together with the part-of-speech tags of the n-1 preceding token)\n",
" + UnigramTagger\n",
" + BigramTagger\n",
" + TrigramTagger"
]
},
{
"cell_type": "code",
"execution_count": 156,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"\"#used from description for RegexpTagger\\npatterns = [(r'.*ing$', 'VBG'), (r'.*ed$', 'VBD'), (r'.*es$', 'VBZ'), (r'.*ould$', 'MD'), (r'.*'s$', 'NN$'), \\n (r'.*s$', 'NNS'), (r'^-?[0-9]+(.[0-9]+)?$', 'CD'), (r'.*', 'NN')]\\n\\n#train taggers\\ndef_model = nltk.DefaultTagger('NN')\\nregexp_model = nltk.RegexpTagger(patterns)\\nuni_model = nltk.UnigramTagger(training_sentences_X1)\\nbi_model = nltk.BigramTagger(training_sentences_X1)\\ntri_model = nltk.TrigramTagger(training_sentences_X1)\""
]
},
"execution_count": 156,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"#see Task 1.3 and 1.6\n",
"\n",
"'''#used from description for RegexpTagger\n",
"patterns = [(r'.*ing$', 'VBG'), (r'.*ed$', 'VBD'), (r'.*es$', 'VBZ'), (r'.*ould$', 'MD'), (r'.*\\'s$', 'NN$'), \n",
" (r'.*s$', 'NNS'), (r'^-?[0-9]+(.[0-9]+)?$', 'CD'), (r'.*', 'NN')]\n",
"\n",
"#train taggers\n",
"def_model = nltk.DefaultTagger('NN')\n",
"regexp_model = nltk.RegexpTagger(patterns)\n",
"uni_model = nltk.UnigramTagger(training_sentences_X1)\n",
"bi_model = nltk.BigramTagger(training_sentences_X1)\n",
"tri_model = nltk.TrigramTagger(training_sentences_X1)'''"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 4. model4 = your POS tagger model (not english)"
]
},
{
"cell_type": "code",
"execution_count": 157,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"#see Task 2.1"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 5. model5 = pre-trained POS tagger model using RDRPOSTagger 1 or TreeTagger 2 (not english)"
]
},
{
"cell_type": "code",
"execution_count": 158,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"#see Task 2.2"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Corpora\n",
"note: data split for training/test = 0.8/0.2 (sequencial)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### 1. X1 = nltk.corpus.treebank (english)"
]
},
{
"cell_type": "code",
"execution_count": 159,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[nltk_data] Downloading package treebank to\n",
"[nltk_data] /Users/Carsten/nltk_data...\n",
"[nltk_data] Package treebank is already up-to-date!\n",
"downloaded treebank\n"
]
}
],
"source": [
"nltk.download('treebank')\n",
"x1 = nltk.corpus.treebank\n",
"print(\"downloaded treebank\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### 2. X2 = nltk.corpus.brown (english)"
]
},
{
"cell_type": "code",
"execution_count": 160,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[nltk_data] Downloading package brown to /Users/Carsten/nltk_data...\n",
"[nltk_data] Package brown is already up-to-date!\n",
"downloaded brown\n"
]
}
],
"source": [
"nltk.download('brown')\n",
"x2 = nltk.corpus.brown\n",
"print(\"downloaded brown\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### 3. X3 = other language German"
]
},
{
"cell_type": "code",
"execution_count": 161,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"#? nltk.corpus.ConllCorpusReader"
]
},
{
"cell_type": "code",
"execution_count": 162,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"downloaded german tiger corpus\n"
]
}
],
"source": [
"# TODO: loading german corpus \n",
"X3 = nltk.corpus.ConllCorpusReader(root='/Users/Carsten/GitRepos/NLP-LAB/Carsten_Solutions/sets/german/', fileids=['tiger_release_aug07.corrected.16012013.conll09'], columntypes=['ignore', 'words', 'ignore', 'ignore', 'pos'], encoding='utf-8')\n",
"german_tagged_sents = X3.tagged_sents()\n",
"print(\"downloaded german tiger corpus\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### 4. X4 = other language Croatian"
]
},
{
"cell_type": "code",
"execution_count": 201,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"#x4 = other language\n",
"#from croatia:\n",
"#by ZˇeljkoAgic ́,⋆NikolaLjubesˇic ́ http://www.lrec-conf.org/proceedings/lrec2014/pdf/690_Paper.pdf\n",
"#licenses: https://creativecommons.org/licenses/by-sa/4.0/\n",
"corp = nltk.corpus.ConllCorpusReader(root=\"/Users/Carsten/GitRepos/NLP-LAB/Carsten_Solutions/sets/croatia/\", fileids=[\"set.hr.conll\"], columntypes=('ignore','words','ignore','ignore','ignore','ignore','ignore','ignore','ignore','pos'))"
]
},
{
"cell_type": "code",
"execution_count": 200,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"<bound method ConllCorpusReader.tagged_sents of <ConllCorpusReader in '/Users/Carsten/GitRepos/NLP-LAB/Carsten_Solutions/sets/croatia'>>\n"
]
}
],
"source": [
"print(corp.tagged_sents.)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Task 1\n",
"* get results for english (plot a graph with all classifiers x results)\n",
" * performance 1.1 = model1 in X1\n",
" * performance 1.2 = model2 in X1\n",
" * performance 1.3.x = model3.x in X1\n",
" * performance 1.4 = model1 in X2\n",
" * performance 1.5 = model2 in X2\n",
" * performance 1.6.x = model3.x in X2"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Generating Testdata"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Generate Training and Testdata for X1\n",
"1. split annotaed sentences into training and testdata\n",
"2. split trainingdata into input data and teacherdata\n",
" *input is the feature vector of each word\n",
" *output is a list of POS tags for each word and sentences"
]
},
{
"cell_type": "code",
"execution_count": 165,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"#to generate trainingsdata, ignore the assigned tags as a function\n",
"def untag(tagged_sentence):\n",
" return [w for w, t in tagged_sentence]"
]
},
{
"cell_type": "code",
"execution_count": 166,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"got 3131 training sentences and 783 test sentences\n"
]
}
],
"source": [
"#print(type(nltk.corpus.treebank.tagged_sents()))\n",
"\n",
"#object including the annotated sentences\n",
"annotated_sent = nltk.corpus.treebank.tagged_sents()\n",
"\n",
"#to split the data, calculate the borders for ratio\n",
"cutoff = int(.8 * len(annotated_sent))\n",
"training_sentences_X1 = annotated_sent[:cutoff]\n",
"test_sentences_X1 = annotated_sent[cutoff:]\n",
"\n",
"#show the amount of sentences\n",
"print(\"got \",len(training_sentences_X1),\" training sentences and \", len(test_sentences_X1), \" test sentences\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"**transform_to_dataset** generates the input X as a list of feature dictinionaries and an output y as a list of pos tags. "
]
},
{
"cell_type": "code",
"execution_count": 167,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"#for training split sentences with its tags into y (for a sentences its resulting tags for each word) and transform sentences and x as a list of the features extracet for echt word in the sentences\n",
"def transform_to_dataset(tagged_sentences):\n",
" X, y = [], []\n",
" for tagged_sentence in tagged_sentences:\n",
" for index in range(len(tagged_sentence)):\n",
" X.append(features(untag(tagged_sentence), index))\n",
" y.append(tagged_sentence[index][1]) \n",
" return X, y"
]
},
{
"cell_type": "code",
"execution_count": 168,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"generated X1 (feature sets) and y1 set of teacher tags\n"
]
}
],
"source": [
"#trainings inputset X and training teacher set y\n",
"X1, y1 = transform_to_dataset(training_sentences_X1)\n",
"print(\"generated X1 (feature sets) and y1 set of teacher tags\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Generate Training and Testdata for X2\n",
"1. split annotaed sentences into training and testdata\n",
"2. split trainingdata into input data and teacherdata\n",
" *input is the feature vector of each word\n",
" *output is a list of POS tags for each word and sentences"
]
},
{
"cell_type": "code",
"execution_count": 169,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"got 45872 training sentences and 11468 test sentences\n"
]
}
],
"source": [
"#object including the annotated sentences\n",
"annotated_sent = nltk.corpus.brown.tagged_sents()\n",
"\n",
"#to split the data, calculate the borders for ratio\n",
"cutoff = int(.8 * len(annotated_sent))\n",
"training_sentences_X2 = annotated_sent[:cutoff]\n",
"test_sentences_X2 = annotated_sent[cutoff:]\n",
"\n",
"#show the amount of sentences\n",
"print(\"got \",len(training_sentences_X2),\" training sentences and \", len(test_sentences_X2), \" test sentences\")"
]
},
{
"cell_type": "code",
"execution_count": 170,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"generated X2 (feature sets) and y2 set of teacher tags\n"
]
}
],
"source": [
"#trainings inputset X and training teacher set y\n",
"X2, y2 = transform_to_dataset(training_sentences_X2)\n",
"print(\"generated X2 (feature sets) and y2 set of teacher tags\")#(X3[:3], y3[:3])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Generate Training and Testdata for X3\n",
"1. split annotaed sentences into training and testdata\n",
"2. split trainingdata into input data and teacherdata\n",
" *input is the feature vector of each word\n",
" *output is a list of POS tags for each word and sentences"
]
},
{
"cell_type": "code",
"execution_count": 171,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"got 40377 training sentences and 10095 test sentences\n"
]
}
],
"source": [
"#object including the annotated sentences\n",
"annotated_sent = X3.tagged_sents()\n",
"\n",
"#print(type(annotated_sent))\n",
"\n",
"#to split the data, calculate the borders for ratio\n",
"cutoff = int(.8 * len(annotated_sent))\n",
"training_sentences_X3 = annotated_sent[:cutoff]\n",
"test_sentences_X3 = annotated_sent[cutoff:]\n",
"\n",
"#show the amount of sentences\n",
"print(\"got \",len(training_sentences_X3),\" training sentences and \", len(test_sentences_X3), \" test sentences\")"
]
},
{
"cell_type": "code",
"execution_count": 172,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"generated X3 (feature sets) and y3 set of teacher tags\n"
]
}
],
"source": [
"#trainings inputset X and training teacher set y\n",
"X3, y3 = transform_to_dataset(training_sentences_X3)\n",
"print(\"generated X3 (feature sets) and y3 set of teacher tags\")#(X3[:3], y3[:3])"
]
},
{
"cell_type": "markdown",
"metadata": {
"collapsed": true
},
"source": [
"#### Implementing a classifier\n",
"relevant imports\n",
"* decision tree as the AI for classfing\n",
"* dict vercorizer transforms the feature dictionary into a vector as the input for the tree"
]
},
{
"cell_type": "code",
"execution_count": 173,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"imported sktree, DictVectorizer, Pipeline\n"
]
}
],
"source": [
"from sklearn.tree import DecisionTreeClassifier\n",
"from sklearn.feature_extraction import DictVectorizer\n",
"from sklearn.pipeline import Pipeline\n",
"print(\"imported sktree, DictVectorizer, Pipeline\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Pipeline manages vectorizer and classifier"
]
},
{
"cell_type": "code",
"execution_count": 174,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Initialized classifier\n"
]
}
],
"source": [
"clf = Pipeline([\n",
" ('vectorizer', DictVectorizer(sparse=False)),\n",
" ('classifier', DecisionTreeClassifier(criterion='entropy'))\n",
"])\n",
"print(\"Initialized classifier\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Calculating performances"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Calculate performance 1.1 - own POS tagger model with X1 = treebank\n",
"* fit the decision tree for a limited amount (size) of training \n",
"* test data and compare with score function on testdata"
]
},
{
"cell_type": "code",
"execution_count": 175,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"training OK\n",
"Accuracy: 0.885273716253\n"
]
}
],
"source": [
"size=10000\n",
"clf.fit(X1[:size], y1[:size])\n",
" \n",
"print('training OK')\n",
" \n",
"X1_test, y1_test = transform_to_dataset(test_sentences_X1)\n",
"\n",
"performance1_1 = clf.score(X1_test, y1_test)\n",
"\n",
"print(\"Accuracy:\", performance1_1)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Calculate performance 1.2 - pre-trained POS tagger model using NLTK (maxentropy english) with X1 = treebank"
]
},
{
"cell_type": "code",
"execution_count": 176,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy: 0.8936074654423873\n"
]
}
],
"source": [
"#extract only the words from feature trainings set\n",
"only_words_X1 = [x['word'] for x in X1_test]\n",
"\n",
"#train with the pos tagger by nltk\n",
"pos_tags_by_pre_trained_pos_tagger = [word_tag_tuple[1] for word_tag_tuple in pos_tag(only_words_X1, lang='eng')]\n",
"\n",
"#calculate performance by comparing each pos tag\n",
"performance1_2 = 0\n",
"for index in range(len(pos_tags_by_pre_trained_pos_tagger)):\n",
" if(pos_tags_by_pre_trained_pos_tagger[index]==y1_test[index]):\n",
" performance1_2 += 1\n",
"performance1_2 /= len(pos_tags_by_pre_trained_pos_tagger)\n",
"\n",
"print(\"Accuracy:\", performance1_2)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Calculate performance 1.3 - rule-based classifiers with X1 = treebank\n",
"1. DefaultTagger that simply tags everything with the same tag\n",
"2. RegexpTagger that applies tags according to a set of regular expressions\n",
"3. N-Gram (n-gram tagger is a generalization of a unigram tagger whose context is the current word together with the part-of-speech tags of the n-1 preceding token)\n",
" + UnigramTagger\n",
" + BigramTagger\n",
" + TrigramTagger"
]
},
{
"cell_type": "code",
"execution_count": 177,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"performance 1.3.1 is: 0.1447677029791906\n",
"performance 1.3.2 is: 0.24232746145017217\n",
"performance 1.3.3 is: 0.8608213982733669\n",
"performance 1.3.4 is: 0.1132791057437996\n",
"performance 1.3.5 is: 0.06736863116922003\n"
]
}
],
"source": [
"#used from description for RegexpTagger\n",
"patterns = [(r'.*ing$', 'VBG'), (r'.*ed$', 'VBD'), (r'.*es$', 'VBZ'), (r'.*ould$', 'MD'), (r'.*\\'s$', 'NN$'), \n",
" (r'.*s$', 'NNS'), (r'^-?[0-9]+(.[0-9]+)?$', 'CD'), (r'.*', 'NN')]\n",
"\n",
"#train taggers\n",
"def_model = nltk.DefaultTagger('NN')\n",
"regexp_model = nltk.RegexpTagger(patterns)\n",
"uni_model = nltk.UnigramTagger(training_sentences_X1)\n",
"bi_model = nltk.BigramTagger(training_sentences_X1)\n",
"tri_model = nltk.TrigramTagger(training_sentences_X1)\n",
"\n",
"#evaluate taggers\n",
"# performance of Default Tagger\n",
"performance1_3_1 = def_model.evaluate(test_sentences_X1)\n",
"print('performance 1.3.1 is: ',performance1_3_1)\n",
"\n",
"# performance of Regex Tagger\n",
"performance1_3_2 = regexp_model.evaluate(test_sentences_X1)\n",
"print('performance 1.3.2 is: ',performance1_3_2)\n",
"\n",
"# performance of Unigram Tagger\n",
"performance1_3_3 = uni_model.evaluate(test_sentences_X1)\n",
"print('performance 1.3.3 is: ',performance1_3_3)\n",
"\n",
"# performance of Bigram Tagger\n",
"performance1_3_4 = bi_model.evaluate(test_sentences_X1)\n",
"print('performance 1.3.4 is: ',performance1_3_4)\n",
"\n",
"# performance of Trigram Tagger\n",
"performance1_3_5 = tri_model.evaluate(test_sentences_X1)\n",
"print('performance 1.3.5 is: ',performance1_3_5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Calculate performance 1.4 - own POS tagger model with X2 = brown"
]
},
{
"cell_type": "code",
"execution_count": 178,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"calculated perfomance 1.4= 0.785464840867\n"
]
}
],
"source": [
"size=10000\n",
"clf.fit(X2[:size], y2[:size])\n",
"X2_test, y2_test = transform_to_dataset(test_sentences_X2)\n",
"performance1_4 = clf.score(X2_test, y2_test)\n",
"print(\"calculated perfomance 1.4= \",performance1_4)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Calculate performance 1.5 - pre-trained POS tagger model using NLTK (maxentropy english) with X2 = brown"
]
},
{
"cell_type": "code",
"execution_count": 179,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy: 0.6044583741861567\n"
]
}
],
"source": [
"#extract only the words from feature trainings set\n",
"only_words_X2 = [x['word'] for x in X2_test]\n",
"\n",
"#train with the pos tagger by nltk\n",
"pos_tags_by_pre_trained_pos_tagger = [word_tag_tuple[1] for word_tag_tuple in pos_tag(only_words_X2, lang='eng')]\n",
"\n",
"#calculate performance by comparing each pos tag\n",
"performance1_5 = 0\n",
"for index in range(len(pos_tags_by_pre_trained_pos_tagger)):\n",
" if(pos_tags_by_pre_trained_pos_tagger[index]==y2_test[index]):\n",
" performance1_5 += 1\n",
"performance1_5 /= len(pos_tags_by_pre_trained_pos_tagger)\n",
"\n",
"print(\"Accuracy:\", performance1_5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Calculate performance 1.6 - rule-based classifiers with X2 = brown"
]
},
{
"cell_type": "code",
"execution_count": 180,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"performance 1.6.1 is: 0.10997763652187324\n",
"performance 1.6.2 is: 0.17594438874995869\n",
"performance 1.6.3 is: 0.8773754310202373\n",
"performance 1.6.4 is: 0.3390490564374869\n",
"performance 1.6.5 is: 0.19178610379738467\n"
]
}
],
"source": [
"uni_model = nltk.UnigramTagger(training_sentences_X2)\n",
"bi_model = nltk.BigramTagger(training_sentences_X2)\n",
"tri_model = nltk.TrigramTagger(training_sentences_X2)\n",
"\n",
"#evaluate taggers\n",
"# performance of Default Tagger\n",
"performance1_6_1 = def_model.evaluate(test_sentences_X2)\n",
"print('performance 1.6.1 is: ',performance1_6_1)\n",
"\n",
"# performance of Regex Tagger\n",
"performance1_6_2 = regexp_model.evaluate(test_sentences_X2)\n",
"print('performance 1.6.2 is: ',performance1_6_2)\n",
"\n",
"# performance of Unigram Tagger\n",
"performance1_6_3 = uni_model.evaluate(test_sentences_X2)\n",
"print('performance 1.6.3 is: ',performance1_6_3)\n",
"\n",
"# performance of Bigram Tagger\n",
"performance1_6_4 = bi_model.evaluate(test_sentences_X2)\n",
"print('performance 1.6.4 is: ',performance1_6_4)\n",
"\n",
"# performance of Trigram Tagger\n",
"performance1_6_5 = tri_model.evaluate(test_sentences_X2)\n",
"print('performance 1.6.5 is: ',performance1_6_5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Using the classifier\n",
"for results the link of pos_tags:\n",
"https://www.ling.upenn.edu/courses/Fall_2003/ling001/penn_treebank_pos.html"
]
},
{
"cell_type": "code",
"execution_count": 181,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"3.6.3\n"
]
}
],
"source": [
"def pos_tag(sentence):\n",
" print('checking...')\n",
" tagged_sentence = []\n",
" tags = clf.predict([features(sentence, index) for index in range(len(sentence))])\n",
" return zip(sentence, tags)\n",
"\n",
"import platform\n",
"print(platform.python_version())\n",
"\n",
"#print(list(pos_tag(word_tokenize('Hello world, lets do something awesome today!'))))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Results for Task 1\n",
"* get results for english (plot a graph with all classifiers x results)\n",
" * performance 1.1 = model1 in X1\n",
" * performance 1.2 = model2 in X1\n",
" * performance 1.3.x = model3.x in X1\n",
" * performance 1.4 = model1 in X2\n",
" * performance 1.5 = model2 in X2\n",
" * performance 1.6.x = model3.x in X2"
]
},
{
"cell_type": "code",
"execution_count": 182,
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"High five! You successfully sent some data to your account on plotly. View your plot in your browser at https://plot.ly/~carsten95/0 or inside your plot.ly account where it is named 'basic-bar'\n"
]
},
{
"data": {
"text/html": [
"<iframe id=\"igraph\" scrolling=\"no\" style=\"border:none;\" seamless=\"seamless\" src=\"https://plot.ly/~carsten95/0.embed\" height=\"525px\" width=\"100%\"></iframe>"
],
"text/plain": [
"<plotly.tools.PlotlyDisplay object>"
]
},
"execution_count": 182,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import plotly\n",
"plotly.tools.set_credentials_file(username='carsten95', api_key='vElf5IOxiFheQdjTxjXW')\n",
"plotly.__version__\n",
"import plotly.plotly as py\n",
"import plotly.graph_objs as go\n",
"\n",
"data = [go.Bar(\n",
" x=['performance 1.1', 'performance 1.2', 'performance 1.3.1', 'performance 1.3.2', 'performance 1.3.3', 'performance 1.3.4', 'performance 1.3.5', 'performance 1.4', 'performance 1.5' , 'performance 1.6.1', 'performance 1.6.2', 'performance 1.6.3', 'performance 1.6.4', 'performance 1.6.5'],\n",
" y=[performance1_1, performance1_2, performance1_3_1, performance1_3_2, performance1_3_3, performance1_3_4, performance1_3_5, performance1_4, performance1_5, performance1_6_1, performance1_6_2, performance1_6_3, performance1_6_4, performance1_6_5]\n",
" )]\n",
"\n",
"py.iplot(data, filename='basic-bar')"
]
},
{
"cell_type": "markdown",
"metadata": {
"collapsed": true
},
"source": [
"### Results for Task 2\n",
"* train your model with standard features (plot a graph with all classifiers x results)\n",
" * performance 2.1 = model4 in X3\n",
" * model 4 your POS tagger model (not english)\n",
" * performance 2.2 = model5 in X3\n",
" * pre-trained POS tagger model using RDRPOSTagger 1 or TreeTagger 2 (not english)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Calculate Performance 2.1"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"again building a pipeline:\n",
"* first vectorizing the dictionary based on feature dict\n",
"* second, initializing and training the max entropy classifier decision tree"
]
},
{
"cell_type": "code",
"execution_count": 183,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"clf = Pipeline([\n",
" ('vectorizer', DictVectorizer(sparse=False)),\n",
" ('classifier', DecisionTreeClassifier(criterion='entropy'))\n",
"])"
]
},
{
"cell_type": "code",
"execution_count": 184,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"training done\n",
"Accuracy: 0.838410907381\n"
]
}
],
"source": [
"size=10000\n",
"clf.fit(X3[:size], y3[:size])\n",
" \n",
"print('training done')\n",
" \n",
"X3_test, y3_test = transform_to_dataset(test_sentences_X3)\n",
"\n",
"performance2_1 = clf.score(X3_test, y3_test)\n",
"\n",
"print(\"Accuracy:\", performance2_1)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Calculate Performance 2.2\n",
"* using RDRPOS Taggger in a python 3 port rom https://github.com/jacopofar/RDRPOSTagger-python-3"
]
},
{
"cell_type": "code",
"execution_count": 185,
"metadata": {},
"outputs": [],
"source": [
"#RDRPOSTagger port python 3 from https://github.com/jacopofar/RDRPOSTagger-python-3"
]
},
{
"cell_type": "code",
"execution_count": 186,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"wrote file de_text.tx in cwd with each word of the sentence seperated by a space\n"
]
}
],
"source": [
"#generate a german txt text file:\n",
"f = open(\"de_text.txt\", 'w')\n",
"for sentence in test_sentences_X3:\n",
" for word, tag in sentence:\n",
" f.write(word + \" \")\n",
" f.write(\"\\n\")\n",
"f.close()\n",
"\n",
"print(\"wrote file de_text.tx in cwd with each word of the sentence seperated by a space\")"
]
},
{
"cell_type": "code",
"execution_count": 187,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"stored: /Users/Carsten/GitRepos/NLP-LAB/Carsten_Solutions\n"
]
}
],
"source": [
"#to use RDRPOSTagger we have to store the path where we are working currently and where the donwnloaded RDRPOSTagger is stored\n",
"import sys, os\n",
"\n",
"#current working directory to restore it later\n",
"dir_path = os.getcwd()\n",
"print(\"stored: \", dir_path)"
]
},
{
"cell_type": "code",
"execution_count": 188,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"switched to path: /Users/Carsten/Downloads/RDRPOSTagger-python-3-master/pSCRDRtagger\n"
]
}
],
"source": [
"#set the rdrpos as path to work in lownloaded api\n",
"RDRPOS_TAGGER_PATH = \"/Users/Carsten/Downloads/RDRPOSTagger-python-3-master/pSCRDRtagger\"\n",
"sys.path.insert(0, RDRPOS_TAGGER_PATH)\n",
"os.chdir(RDRPOS_TAGGER_PATH)\n",
"print(\"switched to path:\", RDRPOS_TAGGER_PATH)"
]
},
{
"cell_type": "code",
"execution_count": 189,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"['Node', '__builtins__', '__cached__', '__doc__', '__file__', '__loader__', '__name__', '__package__', '__spec__', 'tabStr']\n"
]
}
],
"source": [
"# import and rename for easier use\n",
"import RDRPOSTagger\n",
"r = RDRPOSTagger.RDRPOSTagger()\n",
"\n",
"#load files\n",
"r.constructSCRDRtreeFromRDRfile(\"../Models/POS/German.RDR\")\n",
"DICT = RDRPOSTagger.readDictionary(\"../Models/POS/German.DICT\")"
]
},
{
"cell_type": "code",
"execution_count": 190,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"('\\nOutput file:', 'de_text.txt.TAGGED')\n"
]
}
],
"source": [
"#switch back to dir in which we worked at the start\n",
"os.chdir(dir_path)\n",
"\n",
"# generate file with tags after each word with the delimiter /\n",
"r.tagRawCorpus(DICT, \"de_text.txt\")"
]
},
{
"cell_type": "code",
"execution_count": 191,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"#from generated textfile above, seperate the word and tags\n",
"tagged_words = []\n",
"f = open(\"de_text.txt.TAGGED\", 'r')\n",
"for line in f:\n",
" for splits in line.split():\n",
" cmp = splits.rsplit('/',1)\n",
" if len(cmp) != 2:\n",
" print(\"error parsing: \", cmp)\n",
" else:\n",
" w,t = cmp\n",
" tagged_words.append((w,t))"
]
},
{
"cell_type": "code",
"execution_count": 192,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[('CUPERTINO', 'NE'), ('(', '$('), ('rtr', 'NE'), ('/', '$('), ('whp', 'XY'), (')', '$('), ('.', '$.'), ('Der', 'ART'), ('Chef', 'NN'), ('des', 'ART')]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/plain": [
"[[('CUPERTINO', 'NE')], [('(', '$('), ('rtr', 'NE'), ('/', '$('), ('whp', 'XY'), (')', '$('), ('.', '$.')], ...]"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"display(tagged_words[:10])\n",
"display(test_sentences_X3[:10])"
]
},
{
"cell_type": "code",
"execution_count": 193,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy 2.2 = 0.9754407616361072\n"
]
}
],
"source": [
"performance2_2 = 0 # for test \n",
"\n",
"#counter for the words\n",
"i = 0\n",
"\n",
"#evaluate accuracy\n",
"for sent in test_sentences_X3:\n",
" for tagged_w in sent:\n",
" if tagged_w[1] == tagged_words[i][1]:\n",
" performance2_2 += 1\n",
" i += 1\n",
"performance2_2 = performance2_2 / len(tagged_words)\n",
"print(\"Accuracy 2.2 = \",performance2_2)"
]
},
{
"cell_type": "code",
"execution_count": 194,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"High five! You successfully sent some data to your account on plotly. View your plot in your browser at https://plot.ly/~carsten95/0 or inside your plot.ly account where it is named 'basic-bar'\n"
]
},
{
"data": {
"text/html": [
"<iframe id=\"igraph\" scrolling=\"no\" style=\"border:none;\" seamless=\"seamless\" src=\"https://plot.ly/~carsten95/0.embed\" height=\"525px\" width=\"100%\"></iframe>"
],
"text/plain": [
"<plotly.tools.PlotlyDisplay object>"
]
},
"execution_count": 194,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"#visualize results with plotly\n",
"data = [go.Bar(\n",
" x=['performance 2.1', 'performance 2.2'],\n",
" y=[performance2_1, performance2_2]\n",
" )]\n",
"\n",
"py.iplot(data, filename='basic-bar')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"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": "3.6.3"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
Reference in New Issue View Git Blame Copy Permalink