How to group wikipedia categories in python?
You could try to classify the wikipedia categories by the mediawiki links and backlinks returned for each category
import re
from mediawiki import MediaWiki
#TermFind will search through a list a given term
def TermFind(term,termList):
responce=False
for val in termList:
if re.match('(.*)'+term+'(.*)',val):
responce=True
break
return responce
#Find if the links and backlinks lists contains a given term
def BoundedTerm(wikiPage,term):
aList=wikiPage.links
bList=wikiPage.backlinks
responce=False
if TermFind(term,aList)==True and TermFind(term,bList)==True:
responce=True
return responce
container=[]
wikipedia = MediaWiki()
for val in termlist:
cpage=wikipedia.page(val)
if BoundedTerm(cpage,'term')==True:
container.append('medical')
else:
container.append('nonmedical')
The idea is to try to guess a term that is shared by most of the categories, I try biology, medicine and disease with good results. Perhaps you can try to use mulpile calls of BoundedTerms to make the clasification, or a single call for multiple terms and combine the result for the classification. Hope it helps
"Therefore, I would like to know if there is a way to obtain the
parent categoryof the categories (for example, the categories ofenzyme inhibitorandbypass surgerybelong tomedicalparent category)"
MediaWiki categories are themselves wiki pages. A "parent category" is just a category which the "child" category page belongs to. So you can get the parent categories of a category in exactly the same way as you'd obtain the categories of any other wiki page.
For example, using pymediawiki:
p = wikipedia.page('Category:Enzyme inhibitors')
parents = p.categories
Solution Overview
Okay, I would approach the problem from multiple directions. There are some great suggestions here and if I were you I would use an ensemble of those approaches (majority voting, predicting label which is agreed upon by more than 50% of classifiers in your binary case).
I'm thinking about following approaches:
- Active learning (example approach provided by me below)
- MediaWiki backlinks provided as an answer by @TavoGC
- SPARQL ancestral categories provided as a comment to your question by @Stanislav Kralin and/or parent categories provided by @Meena Nagarajan (those two could be an ensemble on their own based on their differences, but for that you would have to contact both creators and compare their results).
This way 2 out of three would have to agree a certain concept is a medical one, which minimizes chance of an error further.
While we're at it I would argue against approach presented by @ananand_v.singh in this answer, because:
- distance metric should not be euclidean, cosine similarity is much better metric (used by, e.g. spaCy) as it does not take into account magnitude of the vectors (and it shouldn't, that's how word2vec or GloVe were trained)
- many artificial clusters would be created if I understood correctly, while we only need two: medicine and non-medicine one. Furthermore, centroid of medicine is not centered on the medicine itself. This poses additional problems, say centroid is moved far away from the medicine and other words like, say,
computerorhuman(or any other not-fitting in your opinion into medicine) might get into the cluster. - it's hard to evaluate results, even more so, the matter is strictly subjective. Furthermore word vectors are hard to visualize and understand (casting them into lower dimensions [2D/3D] using PCA/TSNE/similar for so many words, would give us totally non-sensical results [yeah, I have tried to do it, PCA gets around 5% explained variance for your longer dataset, really, really low]).
Based on the problems highlighted above I have come up with solution using active learning, which is pretty forgotten approach to such problems.
Active Learning approach
In this subset of machine learning, when we have a hard time coming up with an exact algorithm (like what does it mean for a term to be a part of medical category), we ask human "expert" (doesn't actually have to be expert) to provide some answers.
Knowledge encoding
As anand_v.singh pointed out, word vectors are one of the most promising approach and I will use it here as well (differently though, and IMO in a much cleaner and easier fashion).
I'm not going to repeat his points in my answer, so I will add my two cents:
- Do not use contextualized word-embeddings as currently available state of the art (e.g. BERT)
- Check how many of your concepts have no representation (e.g. is represented as a vector of zeros). It should be checked (and is checked in my code,, there will be further discussion when the time comes) and you may use the embedding which has most of them present.
Measuring similarity using spaCy
This class measures similarity between medicine encoded as spaCy's GloVe word vector and every other concept.
class Similarity:
def __init__(self, centroid, nlp, n_threads: int, batch_size: int):
# In our case it will be medicine
self.centroid = centroid
# spaCy's Language model (english), which will be used to return similarity to
# centroid of each concept
self.nlp = nlp
self.n_threads: int = n_threads
self.batch_size: int = batch_size
self.missing: typing.List[int] = []
def __call__(self, concepts):
concepts_similarity = []
# nlp.pipe is faster for many documents and can work in parallel (not blocked by GIL)
for i, concept in enumerate(
self.nlp.pipe(
concepts, n_threads=self.n_threads, batch_size=self.batch_size
)
):
if concept.has_vector:
concepts_similarity.append(self.centroid.similarity(concept))
else:
# If document has no vector, it's assumed to be totally dissimilar to centroid
concepts_similarity.append(-1)
self.missing.append(i)
return np.array(concepts_similarity)
This code will return a number for each concept measuring how similar it is to centroid. Furthermore, it records indices of concepts missing their representation. It might be called like this:
import json
import typing
import numpy as np
import spacy
nlp = spacy.load("en_vectors_web_lg")
centroid = nlp("medicine")
concepts = json.load(open("concepts_new.txt"))
concepts_similarity = Similarity(centroid, nlp, n_threads=-1, batch_size=4096)(
concepts
)
You may substitute you data in place of new_concepts.json.
Look at spacy.load and notice I have used en_vectors_web_lg. It consists of 685.000 unique word vectors (which is a lot), and may work out of the box for your case. You have to download it separately after installing spaCy, more info provided in the links above.
Additionally you may want to use multiple centroid words, e.g. add words like disease or health and average their word vectors. I'm not sure whether that would affect positively your case though.
Other possibility might be to use multiple centroids and calculate similiarity between each concept and multiple of centroids. We may have a few thresholds in such case, this is likely to remove some false positives, but may miss some terms which one could consider to be similar to medicine. Furthermore it would complicate the case much more, but if your results are unsatisfactory you should consider two options above (and only if those are, don't jump into this approach without previous thought).
Now, we have a rough measure of concept's similarity. But what does it mean that a certain concept has 0.1 positive similarity to medicine? Is it a concept one should classify as medical? Or maybe that's too far away already?
Asking expert
To get a threshold (below it terms will be considered non medical), it's easiest to ask a human to classify some of the concepts for us (and that's what active learning is about). Yeah, I know it's a really simple form of active learning, but I would consider it such anyway.
I have written a class with sklearn-like interface asking human to classify concepts until optimal threshold (or maximum number of iterations) is reached.
class ActiveLearner:
def __init__(
self,
concepts,
concepts_similarity,
max_steps: int,
samples: int,
step: float = 0.05,
change_multiplier: float = 0.7,
):
sorting_indices = np.argsort(-concepts_similarity)
self.concepts = concepts[sorting_indices]
self.concepts_similarity = concepts_similarity[sorting_indices]
self.max_steps: int = max_steps
self.samples: int = samples
self.step: float = step
self.change_multiplier: float = change_multiplier
# We don't have to ask experts for the same concepts
self._checked_concepts: typing.Set[int] = set()
# Minimum similarity between vectors is -1
self._min_threshold: float = -1
# Maximum similarity between vectors is 1
self._max_threshold: float = 1
# Let's start from the highest similarity to ensure minimum amount of steps
self.threshold_: float = 1
samplesargument describes how many examples will be shown to an expert during each iteration (it is the maximum, it will return less if samples were already asked for or there is not enough of them to show).steprepresents the drop of threshold (we start at 1 meaning perfect similarity) in each iteration.change_multiplier- if an expert answers concepts are not related (or mostly unrelated, as multiple of them are returned), step is multiplied by this floating point number. It is used to pinpoint exact threshold betweenstepchanges at each iteration.- concepts are sorted based on their similarity (the more similar a concept is, the higher)
Function below asks expert for an opinion and find optimal threshold based on his answers.
def _ask_expert(self, available_concepts_indices):
# Get random concepts (the ones above the threshold)
concepts_to_show = set(
np.random.choice(
available_concepts_indices, len(available_concepts_indices)
).tolist()
)
# Remove those already presented to an expert
concepts_to_show = concepts_to_show - self._checked_concepts
self._checked_concepts.update(concepts_to_show)
# Print message for an expert and concepts to be classified
if concepts_to_show:
print("\nAre those concepts related to medicine?\n")
print(
"\n".join(
f"{i}. {concept}"
for i, concept in enumerate(
self.concepts[list(concepts_to_show)[: self.samples]]
)
),
"\n",
)
return input("[y]es / [n]o / [any]quit ")
return "y"
Example question looks like this:
Are those concepts related to medicine?
0. anesthetic drug
1. child and adolescent psychiatry
2. tertiary care center
3. sex therapy
4. drug design
5. pain disorder
6. psychiatric rehabilitation
7. combined oral contraceptive
8. family practitioner committee
9. cancer family syndrome
10. social psychology
11. drug sale
12. blood system
[y]es / [n]o / [any]quit y
... parsing an answer from expert:
# True - keep asking, False - stop the algorithm
def _parse_expert_decision(self, decision) -> bool:
if decision.lower() == "y":
# You can't go higher as current threshold is related to medicine
self._max_threshold = self.threshold_
if self.threshold_ - self.step < self._min_threshold:
return False
# Lower the threshold
self.threshold_ -= self.step
return True
if decision.lower() == "n":
# You can't got lower than this, as current threshold is not related to medicine already
self._min_threshold = self.threshold_
# Multiply threshold to pinpoint exact spot
self.step *= self.change_multiplier
if self.threshold_ + self.step < self._max_threshold:
return False
# Lower the threshold
self.threshold_ += self.step
return True
return False
And finally whole code code of ActiveLearner, which finds optimal threshold of similiarity accordingly to expert:
class ActiveLearner:
def __init__(
self,
concepts,
concepts_similarity,
samples: int,
max_steps: int,
step: float = 0.05,
change_multiplier: float = 0.7,
):
sorting_indices = np.argsort(-concepts_similarity)
self.concepts = concepts[sorting_indices]
self.concepts_similarity = concepts_similarity[sorting_indices]
self.samples: int = samples
self.max_steps: int = max_steps
self.step: float = step
self.change_multiplier: float = change_multiplier
# We don't have to ask experts for the same concepts
self._checked_concepts: typing.Set[int] = set()
# Minimum similarity between vectors is -1
self._min_threshold: float = -1
# Maximum similarity between vectors is 1
self._max_threshold: float = 1
# Let's start from the highest similarity to ensure minimum amount of steps
self.threshold_: float = 1
def _ask_expert(self, available_concepts_indices):
# Get random concepts (the ones above the threshold)
concepts_to_show = set(
np.random.choice(
available_concepts_indices, len(available_concepts_indices)
).tolist()
)
# Remove those already presented to an expert
concepts_to_show = concepts_to_show - self._checked_concepts
self._checked_concepts.update(concepts_to_show)
# Print message for an expert and concepts to be classified
if concepts_to_show:
print("\nAre those concepts related to medicine?\n")
print(
"\n".join(
f"{i}. {concept}"
for i, concept in enumerate(
self.concepts[list(concepts_to_show)[: self.samples]]
)
),
"\n",
)
return input("[y]es / [n]o / [any]quit ")
return "y"
# True - keep asking, False - stop the algorithm
def _parse_expert_decision(self, decision) -> bool:
if decision.lower() == "y":
# You can't go higher as current threshold is related to medicine
self._max_threshold = self.threshold_
if self.threshold_ - self.step < self._min_threshold:
return False
# Lower the threshold
self.threshold_ -= self.step
return True
if decision.lower() == "n":
# You can't got lower than this, as current threshold is not related to medicine already
self._min_threshold = self.threshold_
# Multiply threshold to pinpoint exact spot
self.step *= self.change_multiplier
if self.threshold_ + self.step < self._max_threshold:
return False
# Lower the threshold
self.threshold_ += self.step
return True
return False
def fit(self):
for _ in range(self.max_steps):
available_concepts_indices = np.nonzero(
self.concepts_similarity >= self.threshold_
)[0]
if available_concepts_indices.size != 0:
decision = self._ask_expert(available_concepts_indices)
if not self._parse_expert_decision(decision):
break
else:
self.threshold_ -= self.step
return self
All in all, you would have to answer some questions manually but this approach is way more accurate in my opinion.
Furthermore, you don't have to go through all of the samples, just a small subset of it. You may decide how many samples constitute a medical term (whether 40 medical samples and 10 non-medical samples shown, should still be considered medical?), which let's you fine-tune this approach to your preferences. If there is an outlier (say, 1 sample out of 50 is non-medical), I would consider the threshold to still be valid.
Once again: This approach should be mixed with others in order to minimalize the chance for wrong classification.
Classifier
When we obtain the threshold from expert, classification would be instantenous, here is a simple class for classification:
class Classifier:
def __init__(self, centroid, threshold: float):
self.centroid = centroid
self.threshold: float = threshold
def predict(self, concepts_pipe):
predictions = []
for concept in concepts_pipe:
predictions.append(self.centroid.similarity(concept) > self.threshold)
return predictions
And for brevity, here is the final source code:
import json
import typing
import numpy as np
import spacy
class Similarity:
def __init__(self, centroid, nlp, n_threads: int, batch_size: int):
# In our case it will be medicine
self.centroid = centroid
# spaCy's Language model (english), which will be used to return similarity to
# centroid of each concept
self.nlp = nlp
self.n_threads: int = n_threads
self.batch_size: int = batch_size
self.missing: typing.List[int] = []
def __call__(self, concepts):
concepts_similarity = []
# nlp.pipe is faster for many documents and can work in parallel (not blocked by GIL)
for i, concept in enumerate(
self.nlp.pipe(
concepts, n_threads=self.n_threads, batch_size=self.batch_size
)
):
if concept.has_vector:
concepts_similarity.append(self.centroid.similarity(concept))
else:
# If document has no vector, it's assumed to be totally dissimilar to centroid
concepts_similarity.append(-1)
self.missing.append(i)
return np.array(concepts_similarity)
class ActiveLearner:
def __init__(
self,
concepts,
concepts_similarity,
samples: int,
max_steps: int,
step: float = 0.05,
change_multiplier: float = 0.7,
):
sorting_indices = np.argsort(-concepts_similarity)
self.concepts = concepts[sorting_indices]
self.concepts_similarity = concepts_similarity[sorting_indices]
self.samples: int = samples
self.max_steps: int = max_steps
self.step: float = step
self.change_multiplier: float = change_multiplier
# We don't have to ask experts for the same concepts
self._checked_concepts: typing.Set[int] = set()
# Minimum similarity between vectors is -1
self._min_threshold: float = -1
# Maximum similarity between vectors is 1
self._max_threshold: float = 1
# Let's start from the highest similarity to ensure minimum amount of steps
self.threshold_: float = 1
def _ask_expert(self, available_concepts_indices):
# Get random concepts (the ones above the threshold)
concepts_to_show = set(
np.random.choice(
available_concepts_indices, len(available_concepts_indices)
).tolist()
)
# Remove those already presented to an expert
concepts_to_show = concepts_to_show - self._checked_concepts
self._checked_concepts.update(concepts_to_show)
# Print message for an expert and concepts to be classified
if concepts_to_show:
print("\nAre those concepts related to medicine?\n")
print(
"\n".join(
f"{i}. {concept}"
for i, concept in enumerate(
self.concepts[list(concepts_to_show)[: self.samples]]
)
),
"\n",
)
return input("[y]es / [n]o / [any]quit ")
return "y"
# True - keep asking, False - stop the algorithm
def _parse_expert_decision(self, decision) -> bool:
if decision.lower() == "y":
# You can't go higher as current threshold is related to medicine
self._max_threshold = self.threshold_
if self.threshold_ - self.step < self._min_threshold:
return False
# Lower the threshold
self.threshold_ -= self.step
return True
if decision.lower() == "n":
# You can't got lower than this, as current threshold is not related to medicine already
self._min_threshold = self.threshold_
# Multiply threshold to pinpoint exact spot
self.step *= self.change_multiplier
if self.threshold_ + self.step < self._max_threshold:
return False
# Lower the threshold
self.threshold_ += self.step
return True
return False
def fit(self):
for _ in range(self.max_steps):
available_concepts_indices = np.nonzero(
self.concepts_similarity >= self.threshold_
)[0]
if available_concepts_indices.size != 0:
decision = self._ask_expert(available_concepts_indices)
if not self._parse_expert_decision(decision):
break
else:
self.threshold_ -= self.step
return self
class Classifier:
def __init__(self, centroid, threshold: float):
self.centroid = centroid
self.threshold: float = threshold
def predict(self, concepts_pipe):
predictions = []
for concept in concepts_pipe:
predictions.append(self.centroid.similarity(concept) > self.threshold)
return predictions
if __name__ == "__main__":
nlp = spacy.load("en_vectors_web_lg")
centroid = nlp("medicine")
concepts = json.load(open("concepts_new.txt"))
concepts_similarity = Similarity(centroid, nlp, n_threads=-1, batch_size=4096)(
concepts
)
learner = ActiveLearner(
np.array(concepts), concepts_similarity, samples=20, max_steps=50
).fit()
print(f"Found threshold {learner.threshold_}\n")
classifier = Classifier(centroid, learner.threshold_)
pipe = nlp.pipe(concepts, n_threads=-1, batch_size=4096)
predictions = classifier.predict(pipe)
print(
"\n".join(
f"{concept}: {label}"
for concept, label in zip(concepts[20:40], predictions[20:40])
)
)
After answering some questions, with threshold 0.1 (everything between [-1, 0.1) is considered non-medical, while [0.1, 1] is considered medical) I got the following results:
kartagener s syndrome: True
summer season: True
taq: False
atypical neuroleptic: True
anterior cingulate: False
acute respiratory distress syndrome: True
circularity: False
mutase: False
adrenergic blocking drug: True
systematic desensitization: True
the turning point: True
9l: False
pyridazine: False
bisoprolol: False
trq: False
propylhexedrine: False
type 18: True
darpp 32: False
rickettsia conorii: False
sport shoe: True
As you can see this approach is far from perfect, so the last section described possible improvements:
Possible improvements
As mentioned in the beginning using my approach mixed with other answers would probably leave out ideas like sport shoe belonging to medicine out and active learning approach would be more of a decisive vote in case of a draw between two heuristics mentioned above.
We could create an active learning ensemble as well. Instead of one threshold, say 0.1, we would use multiple of them (either increasing or decreasing), let's say those are 0.1, 0.2, 0.3, 0.4, 0.5.
Let's say sport shoe gets, for each threshold it's respective True/False like this:
True True False False False,
Making a majority voting we would mark it non-medical by 3 out of 2 votes. Furthermore, too strict threshold would me mitigated as well if thresholds below it out-vote it (case if True/False would look like this: True True True False False).
Final possible improvement I came up with: In the code above I'm using Doc vector, which is a mean of word vectors creating the concept. Say one word is missing (vectors consisting of zeros), in such case, it would be pushed further away from medicine centroid. You may not want that (as some niche medical terms [abbreviations like gpv or others] might be missing their representation), in such case you could average only those vectors which are different from zero.
I know this post is quite lengthy, so if you have any questions post them below.
There is a concept of word Vectors in NLP, what it basically does is by looking through mass volumes of text, it tries to convert words to multi-dimensional vectors and then lesser the distance between those vectors, greater the similarity between them, the good thing is that many people have already generated this word vectors and made them available under very permissive licences, and in your case you are working with Wikipedia and there exist word vectors for them here http://dumps.wikimedia.org/enwiki/latest/enwiki-latest-pages-articles.xml.bz2
Now these would be the most suited for this task since they contain most words from Wikipedia's corpora, but in case they are not suited for you, or are removed in the future you can use one from I will list below more of these, with that said, there is a better way to do this, i.e. by passing them to tensorflow's universal language model embed module in which you don't have to do most of the heavy lifting, you can read more about that here. The reason I put it after the Wikipedia text dump is because I have heard people say that they are a bit hard to work with when working with medical samples. This paper does propose a solution to tackle that but I have never tried that so I cannot be sure of it's accuracies.
Now how you can use the word embeddings from tensorflow is simple, just do
embed = hub.Module("https://tfhub.dev/google/universal-sentence-encoder/2")
embeddings = embed(["Input Text here as"," List of strings"])
session.run(embeddings)
Since you might not be familiar with tensorflow and trying to run just this piece of code you might run into some troubles, Follow this link where they have mentioned completely how to use this and from there you should be able to easily modify this to your needs.
With that said I would recommend first checking out he tensorlfow's embed module and their pre-trained word embedding's, if they don't work for you check out the Wikimedia link, if that also doesn't work then proceed to the concepts of the paper I have linked. Since this answer is describing an NLP approach, it will not be 100% accurate, so keep that in mind before you proceed.
Glove Vectors https://nlp.stanford.edu/projects/glove/
Facebook's fast text: https://github.com/facebookresearch/fastText/blob/master/pretrained-vectors.md
Or this http://www.statmt.org/lm-benchmark/1-billion-word-language-modeling-benchmark-r13output.tar.gz
If you run into problems implementing this after following the colab tutorial add your problem to the question and comment below, from there we can proceed further.
Edit Added code to cluster topics
Brief, Rather than using words vector, I am encoding their summary sentences
file content.py
def AllTopics():
topics = []# list all your topics, not added here for space restricitons
for i in range(len(topics)-1):
yield topics[i]
File summaryGenerator.py
import wikipedia
import pickle
from content import Alltopics
summary = []
failed = []
for topic in Alltopics():
try:
summary.append(wikipedia.summary(tuple((topic,str(topic)))))
except Exception as e:
failed.append(tuple((topic,e)))
with open("summary.txt", "wb") as fp:
pickle.dump(summary , fp)
with open('failed.txt', 'wb') as fp:
pickle.dump('failed', fp)
File SimilartiyCalculator.py
import tensorflow as tf
import tensorflow_hub as hub
import numpy as np
import os
import pandas as pd
import re
import pickle
import sys
from sklearn.cluster import AgglomerativeClustering
from sklearn import metrics
from scipy.cluster import hierarchy
from scipy.spatial import distance_matrix
try:
with open("summary.txt", "rb") as fp: # Unpickling
summary = pickle.load(fp)
except Exception as e:
print ('Cannot load the summary file, Please make sure that it exists, if not run Summary Generator first', e)
sys.exit('Read the error message')
module_url = "https://tfhub.dev/google/universal-sentence-encoder-large/3"
embed = hub.Module(module_url)
tf.logging.set_verbosity(tf.logging.ERROR)
messages = [x[1] for x in summary]
labels = [x[0] for x in summary]
with tf.Session() as session:
session.run([tf.global_variables_initializer(), tf.tables_initializer()])
message_embeddings = session.run(embed(messages)) # In message embeddings each vector is a second (1,512 vector) and is numpy.ndarray (noOfElemnts, 512)
X = message_embeddings
agl = AgglomerativeClustering(n_clusters=5, affinity='euclidean', memory=None, connectivity=None, compute_full_tree='auto', linkage='ward', pooling_func='deprecated')
agl.fit(X)
dist_matrix = distance_matrix(X,X)
Z = hierarchy.linkage(dist_matrix, 'complete')
dendro = hierarchy.dendrogram(Z)
cluster_labels = agl.labels_
This is also hosted on GitHub at https://github.com/anandvsingh/WikipediaSimilarity Where you can find the similarity.txt file, and other files, In my case I couldn't run it on all the topics, but I would urge you to run it on the full list of topics (Directly clone the repository and run SummaryGenerator.py), and upload the similarity.txt via a pull request in case you don't get expected result. And if possible also upload the message_embeddings in a csv file as topics and there embeddings.
Changes after edit 2
Switched the similarityGenerator to a hierarchy based clustering(Agglomerative) I would suggest you to keep the title names at the bottom of the dendrogram and for that look at the definition of dendrogram here, I verified viewing some samples and the results look quite good, you can change the n_clusters value to fine tune your model. Note: This requires you to run summary generator again. I think you should be able to take it from here, what you have to do is try a few values of n_cluster and see in which all medical terms are grouped together, then find the cluster_label for that cluster and you are done. Since here we group by summary, the clusters will be more accurate. If you run into any problems or don't understand something, comment below.