I am currently breaking my head over a problem concerning matching of a dataset containing self reported job titles (and thus messy, incoherent and with loads of orthographic errors) with a standardised list of official titles in SAS.
In a condensed form lets say I have the following list of standardised titles:
machine engineer
machine assistant
machine mechanic
machine operator
And a snapshot of the self-reported titles with some of the common problems:
Machine engineer
machine engineer at ABC machine company
mechanic for agricultural machines
mchaine assistant
machine operator/conductor
Often these self-reported titles contain a company name and a sector that is not of interest as well as errors in spelling. First I was thinking of a fuzzy matching, but with the COMPGED function the edit distance would be very high for the entries that for example contain the company. The problem itself contains approximately 1,400 standardised job titles and more than 170,000 distinct self-reported titles coming from a total of roughly 1 million entries. Obviously it would be too optimistic to expect that all titles could be matched, but any method of getting closer would be of great help. What I am aiming for would essentially look like this:
ID self_reported standardised
1 Machine engineer machine engineer
2 machine engineer at ABC machine company machine engineer
3 mechanic for agricultural machines machine mechanic
4 mchaine assistant machine assistant
5 machine operator/conductor machine operator
Is there any method that is able to somewhat deal with the multitude of problems that can arise in this matching problem?
Thank you!
Related
Im struggling to understand how to make an inventory application:
Store inventory data (item name, sku, descriptions, prices, inventory counts)
Allow computers access from different networks
Keep computers updated. By that I dont understand how they communicate to each other across different networks/connections, so that if one computer makes a change to say inventory count how does it let the other computers know to subtract say 5 of an item bc 5 were sold.
question 1 I can understand how storage works to a degree, but when it comes to 2 and 3 how would you approach it? I figured servers would be the answer but then the question becomes what kind? AWS RDS? If you have links to tutorials that can explain this process and different approaches please feel free to link
Ive been trying to figure this out for a few months now and I think im getting more confused than answers (I swear I get more inventory application ADs crap than actual information, like no joke try searching "how inventory applications store and update data and keep computers updated" or any phrase like that... and its a bazillion freaking ads lol)
I am trying to implement something similar in https://arxiv.org/pdf/1603.04259.pdf using awesome gensim library however I am having trouble improving quality of results when I compare to Collaborative Filtering.
I have two models one built on Apache Spark and other one using gensim Word2Vec on grouplens 20 million ratings dataset. My apache spark model is hosted on AWS http://sparkmovierecommender.us-east-1.elasticbeanstalk.com
and I am running gensim model on my local. However when I compare the results I see superior results with CF model 9 out of 10 times(like below example more similar to searched movie - affinity towards Marvel movies)
e.g.:- If I search for Thor movie I get below results
Gensim
Captain America: The First Avenger (2011)
X-Men: First Class (2011)
Rise of the Planet of the Apes (2011)
Iron Man 2 (2010)
X-Men Origins: Wolverine (2009)
Green Lantern (2011)
Super 8 (2011)
Tron:Legacy (2010)
Transformers: Dark of the Moon (2011)
CF
Captain America: The First Avenger
Iron Man 2
Thor: The Dark World
Iron Man
The Avengers
X-Men: First Class
Iron Man 3
Star Trek
Captain America: The Winter Soldier
Below is my model configuration, so far I have tried playing with window, min_count and size parameter but not much improvement.
word2vec_model = gensim.models.Word2Vec(
seed=1,
size=100,
min_count=50,
window=30)
word2vec_model.train(movie_list, total_examples=len(movie_list), epochs=10)
Any help in this regard is appreciated.
You don't mention what Collaborative Filtering algorithm you're trying, but maybe it's just better than Word2Vec for this purpose. (Word2Vec is not doing awful; why do you expect it to be better?)
Alternate meta-parameters might do better.
For example, the window is the max-distance between tokens that might affect each other, but the effective windows used in each target-token training randomly chosen from 1 to window, as a way to give nearby tokens more weight. Thus when some training-texts are much larger than the window (as in your example row), some of the correlations will be ignored (or underweighted). Unless ordering is very significant, a giant window (MAX_INT?) might do better, or even a related method where ordering is irrelevant (such as Doc2Vec in pure PV-DBOW dm=0 mode, with every token used as a doc-tag).
Depending on how much data you have, your size might be too large or small. Different min_count, negative count, greater 'iter'/'epochs', or sample level might work much better. (And perhaps even things you've already tinkered with would only help after other changes are in place.)
I am trying to predict match winner based on the historical data set as shown below,
The data set comprises of IPL seasons and Team_Name_id vs Opponent Team are the team names in IPL. I have set the match id as Row id and created the model. When running realtime testing, the result is not as expected (shown below)
Target is set as Match_winner_id.
Am I missing any configurations? Please help
The model is working perfectly correctly. There's just two problems:
Your input data is not very good
There's no way for the model to know that only one of those two teams should win
Data Quality
A predictive model needs good quality input data on which to reverse-engineer a model that explains a given result. This input data should contain information that can be used to predict a result given a different set of input data.
For example, when predicting house prices, it would need to know the suburb (category), number of bedrooms/bathrooms/parking spaces, age of the building and selling price. It could then predict the selling price for other houses with a slightly different mix of variables.
However, based on your screenshot, you are giving the following information (and probably more) on which to make your prediction:
Teams: Not great, because you are separating Column C and Column D. The model will assume they are unrelated information. It doesn't realise that those two values could be swapped.
Match date: Useless information unless the outcome varies in proportion to time (eg a team continually gets better)
Season: As with Match Date, this is probably useless because you're always predicting the future -- you won't be predicting for a past season
Venue: Only relevant if a particular team always wins at a given venue
Toss Decision: Would this really influence the outcome? Also, it's only known once the game begins, so not great for predicting a future game.
Win Type: You won't know the win type until a game is over, so it's not suitable for predicting a future game.
Score: Again, not known until the actual game, so no good for future predictions.
Man of the Match: Not known for future games.
Umpire: How does an umpire influence the result of a game?
City: Yes, given that home teams often have an advantage.
You have provided very little information that could be used to predict a future game. There is really only the teams and the venue. Everything else is either part of the game itself or irrelevant.
Picking only one of the two teams
When the ML model looks at your data and tries to make a prediction, it will look at all the data you have provided. For example, it might notice that for a given venue and season, Team 8 has a higher propensity to win. Therefore, given that venue and season, it will favour a win by Team 8. The model has no concept that the only possible outcome is one of the two teams given in columns C and D.
You are predicting for two given teams and you are listing the teams in either Column C or Column D and this makes no sense -- the result is the same if you swapped the teams between columns, but the model has no concept of this. Also, information about Team 1 vs Team 2 is totally irrelevant for Team 3 vs Team 4.
What you should do is create one dataset per team, listing all their matches, plus a column that shows the outcome -- either a boolean (Win/Lose) or a value that represents the number of runs by which they won (where negative is a loss). You would then ask them model to predict the result for that team, given the input data, which would be win/lose or a points above/below the other team.
But at the core, I think that your input data doesn't have enough rich content to be able to make a sensible prediction. Just ask yourself: "What data would I like to know if I were to guess which team would win?" It would probably be past results, weather conditions, which players were on each team, how many matches they played in the last week, etc. None of this information is being provided as input on each line of your input data.
I'm in the process of (re-) training spaCy's Named Entity Recognizer and have a couple of doubts that I hope a more experienced researcher/practitioner can help me figure out:
If a few hundred examples are considered 'a good starting point', then what would be a reasonable number to aim for? Is 100 000 entity/label excessive?
If I introduce a new label, is it best if the number of the entities of that labeled are roughly the same (balanced) during training?
Regarding the mixing in 'examples of other entity types':
do I just add random known categories/labels to my training set eg: ('The Business Standard published in its recent issue on crude oil and natural gas ...', [(4,21, 'ORG')], )?
can I use the same text for various labels? e.g. ('The Business Standard published in its recent issue on crude oil and natural gas ...', [(55,64, 'COMMODITY')], )?
on a similar note let's assume I want spaCyto also recognize a second COMMODITY could I then just use the same sentence and label a different region e.g. ('The Business Standard published in its recent issue on crude oil and natural gas ...', [(69,80, 'COMMODITY')], )? Is that how it's supposed to be done?
what ratio between new and other (old) labels is considered reasonable
Thanks
PS I'm working with Python2.7 in Ubuntu 16.04 using spaCy 1.8.2
For a full answer by Matthew Honnibal check out issue 1054 on spaCy's github page. Below are the most important points as they relate to my questions:
Question(Q) 1: If a few hundred examples are considered 'a good starting point', then what would be a reasonable number to aim for? Is 100 000 entity/label excessive?
Answer(A): Every machine learning problem will have a different examples/accuracy curve. You can get an idea for this by training with less data than you have, and seeing what the curve looks like. If you have 1,000 examples, then try training with 500, 750, etc, and see how that affects your accuracy.
Q 2: If I introduce a new label, is it best if the number of the entities of that label are roughly the same (balanced) during training?
A: There's trade-off between making the gradients too sparse, and making the learning problem too unrepresentative of what the actual examples will look like.
Q 3: Regarding the mixing in 'examples of other entity types':
do I just add random known categories/labels to my training set:
A: No, one should annotate all the entities in that text, so the example above: ('The Business Standard published in its recent issue on crude oil and natural gas ...', [(4,21, 'ORG')], ) should be ('The Business Standard published in its recent issue on crude oil and natural gas ...', [(4,21, 'ORG'), (55,64, 'COMMODITY'), (69,80, 'COMMODITY')], )
can I use the same text for various labels?:
A: Not in the way the examples were given. See previous answer.
what ratio between new and other (old) labels is considered reasonable?:
A: See answer Q 2.
PS: Double citations are direct quotes from the github issue answer.
Is there a web service or a tool to detect if what a certain text is the name or a person, a place or an object (device).
eg:
Input: Bill Clinton Output: Person
Input: Blackberry Output: Device
Input: New york Output: Place
Accuracy can be low. I have looked at opencyc but I couldnt get it to work. Is there a way I can use WikiPedia for this?
For a start separating a person or a thing will be great.
I think wikipedia would be a very good source. Given the input, you could try and find an entry in wikipedia and scrape the resulting page (if it exists).
Persons and Places should have fairly distinct sets of data - birthdates, locations, etc in the article that you could use to tell them apart, and anything else is an object.
It's worth a shot anyway.
Looking at the output of Wolfram Alpha, it seems that you can possibly identify a person by searching Bill Clinton Birthday or just Bill Clinton, or you can identify a location by searching New York GPS coordinates or just New York, for even better results. Blackberry seems like a tough word for Alpha, because it keeps wanting to interpret it as a fruit. You might have luck searching Froogle to identify a device.
It seems like WA will give you a fairly decent accuracy, at least if you're using famous people/places.
How about using a search engine? Google would be good, and I think Yahoo! has tools for building your own search.
I googled:
Results 1 - 10 of about 27,100,000 for "bill clinton" person
Results 1 - 10 of about 6,050,000 for "bill clinton" place
Results 1 - 10 of about 601,000 for "bill clinton" device
He's a person!
Results 1 - 10 of about 391,000,000 for "new york" place.
Results 1 - 10 of about 280,000,000 for "new york" person.
Results 1 - 10 of about 84,100,000 for "new york" device.
It's a place!
Results 1 - 10 of about 11,000,000 for "blackberry" person
Results 1 - 10 of about 36,600,000 for "blackberry" place
Results 1 - 10 of about 28,000,000 for "blackberry" device
Unfortunately, blackberry is a place as well. :-/
Note that only in the case of 'blackberry' did "device" even get close. Maybe you need to weight the page hit values. What is your application? Do you have any idea which "devices" you'd have to classify? What is the possible range of inputs?
Maybe you want to combine the results you get from different sources.
I think the basic task you're trying to accomplish is more formally known as named entity recognition. This task is nontrivial, and by only inputting the name stripped of any context, you're making it even harder.
For example, we'd like to think examples such as "Bill Clinton" and "New York" are obviously unambiguous, but looking at their disambiguation pages in Wikipedia shows that there are several potential entities they may refer to. "New York" is both a state, city, and movie title. "Bill Clinton" is a bit less ambiguous if you're only looking at Wikipedia, but I'm sure you'll find dozens of Bill Clintons in any phonebook. It might also be the name of someone's sailboat or pet dog. What if someone inputs "Washington"? That could be both a U.S. President, state, district, city, lake, street, island, movie, one of several U.S. navy ships, bridge, as well as other things. Determining which is the "correct" usage you'd want the webservice to return could become very complicated.
As much as Cyc knows, I think you'll find it's still not as comprehensive as Wikipedia. However, the main downside to Wikipedia is that it's essentially unstructured. Personally, I find Cyc's API so convoluted and poorly documented, that parsing Wikipedia's natural language almost seems easier.
If I had to implement such a webservice from scratch, I'd start by downloading a snapshot of Wikipedia, and then writing a parser that would read through all the articles, and generate a named entity index based on article titles. You could manually "classify" a few dozen examples as person/place/object, and train a classifier (Bayesian,Maxent,SVM) to automatically classify other examples based on the word frequencies of their articles.