scratchpad

http://www.grcdi.nl/gsb/summary_%20company%20legal%20forms.html
https://en.wikipedia.org/wiki/Types_of_business_entity

https://gate.ac.uk/sale/tao/splitap6.html
https://github.com/hpcc-systems/TextAnalytics/blob/master/hpcc/GATE_Annie/plugins/ANNIE/resources/NE/name.jape


## Java * References
https://www.oreilly.com/learning/java-8-functional-interfaces
http://refactoring.info/tools/LambdaFicator/
https://github.com/JnRouvignac/AutoRefactor/wiki/Useful-links

https://mailparser.io/pricing

https://parser.zapier.com/
http://unitedobjectives.com/products/invoice_processor/

InvoiceP2 is built in Python and C (for performance critical modules). It uses proprietary Natural Language and Text Processing technology created by United Objectives. Parsing algorithms are backed up by semantic, frequency, colocation, proper names and morphological dictionaries that enable the system to parse documents without any predefined structure.

CloudFactory is a scalable way to outsource tedious and repetitive data work. We help break your project down into small tasks that are processed by our global 24x7 workforce. Our software platform manages this workforce to ensure quick turnaround and accurate results for your business.

http://www.raremile.com/mobile-receipt-scanning-and-data-extraction.html


Papers
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.403.1594&rep=rep1&type=pdf
http://vision.gyte.edu.tr/publications/2016/VISIGRAPP_2016Camera.pdf
http://www.ict.griffith.edu.au/das2012/attachments/FullPaperProceedings/4661a409.pdf
http://www.dsi.unifi.it/~simone/Papers/ICDAR97b.pdf
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.143.137&rep=rep1&type=pdf
http://ocean.kisti.re.kr/downfile/volume/kmtd/MTMDCW/2010/v13n12/MTMDCW_2010_v13n12_1786.pdf
https://www.researchgate.net/profile/Enrico_Francesconi/publication/3710626_Rectangle_labelling_for_an_invoice_understanding_system/links/02e7e51818f35768f6000000.pdf
http://seco.cs.aalto.fi/u/jwtuomin/svn/secoweb/public_html/publications/2011/nyberg-masters-thesis.pdf
https://jamia.oxfordjournals.org/content/21/5/850


paper
Converting and Annotating Quantitative Data Tables
https://files.ifi.uzh.ch/ddis/iswc_archive/iswc/ab/2010/iswc2010.semanticweb.org/pdf/167.pdf

Rule Based Systems for Big Data
http://www.springer.com/us/book/9783319236957

GSearch Queries:
annotation of quantitative research data stored in tables.
table header disambiguation

http://iswc2013.semanticweb.org/sites/default/files/iswc_poster_7.pdf
Towards Disambiguating Web Tables

create a file with simple Table
GUI
create a application
- annotate it with tokens
- create the jape file for it and test it output
- save as the applicataion
library
- load the application from file
- execute the PR
print out

- the reason mail parsing is slow because its following rules based parsing on labels. Matching each text against some list of patterns
- instead we should use the model based parsing where probabilistic output is provided

We can also use a combination:
- use the rule based parser to prepare the learning set for the probabilistic parser
- use the rule parser in combination with the ML

Currently I am doing following steps :

https://gate.ac.uk/wiki/TrainingCourseJune2015/

jape training
https://gate.ac.uk/sale/talks/gate-course-jun13/track-1/module-3-jape/module-3-jape.pdf

jake wiki
https://gate.ac.uk/wiki/jape-repository/

GATE Training docs
https://gate.ac.uk/wiki/TrainingCourseJune2015/

"","SportsInterest","MoviesInterest","Technology.Interest","Finance.Interest","Politics.Interest","Travel.Interest","BizInterest","Intnl     Interest","Age","Gender.Female","Gender.Male","Relationship.Status.Divorced","Relationship.Status.Married","Relationship.Status.Single",     "Family.Size","Job.Level.Director","Job.Level.Entry.Level","Job.Level.intern","Job.Level.Manager","Job.Level.Sr..Manager","Income.25.to.     35","Income.35.to.50","Income.50.to.75","Income.75.to.100","Income.100.plus","Income.25.minus","No.of.Vehicles.Owned","Sales"