Similar String Matching Algorithm : Jaccard, Dice ,Cosine and Exact - string

I am trying to find out the similarity between two english sentences. Among Jaccard, Dice, Exact and Cosine string matching algorithm which is the best when it comes to string matching or determining the closeness?
Sentence 1: Online shopping for electronics, computer parts, apple accessories, health & beauty, video games, cell phone accessories, home & garden and more at tmart.com. we provide wide selections of products at best price for worldwide free shipping.
Sentence 2: Shop for electronics, apparels & more using our Flipkart app Free shipping & COD.

Related

EXCEL: Look for cells in column with fuzzy duplicates

my problem is as follows:
Imagine I have a large Excel dataset that has a company identifier in column (A) and long string texts within cells in column (B). These texts in those cells contain many words, imagine it as being the company description. I need to find fuzzy duplicates in this database, i.e. an example could be:
In cell B2 I have:
"Amazon.com Inc (Amazon) is an online retailer and web service provider. The company provides products such as apparel, auto and industrial items, beauty and health products, electronics, grocery, books, games, jewellery, kids and baby products, movies, music, sports goods, toys, tools and other related products."
and in cell B222 I have:
"%& COMPANY DESCRIPTION: Amazon is an online retailer and web service provider. Amazon provides products such as apparel, auto and industrial items, beauty and health products, electronics, grocery, books, games, jewellery, kids and baby products, movies, music, sports goods, toys, tools and other related products. Amazon is a great company."
So my point is: is there a way to find B222 fast and somehow show in B2 that there is a fuzzy duplicate, e.g. with an 80% match, in B222?
I have tried multiple tools such as Ablebits and the Levenshtein Distance in VBA. However I am not 100% satisfied with the result.
Thank you for any help!
Best,

Excel COUNTIF for comma separated values

I have the following range of values:
National museum of Natural History
Archaeological Museum, Art Museum, Agricultural Museum, Marine Museum
National museum of Natural History, Art Museum
National museum of Natural History, Archaeological Museum, Science Museum
Open-air Museum, Art Museum
This means there can be more than one value in one column, separated by commas.
I want to count each value, for example: all "Archaeological Museum" in this range. I have used the COUNTIF method but it seems not to be enough:
COUNTIF(A1:A6, "Archaeological Museum")
=> result: 0
=> expected result: 2
Is there an only way to spilt the column based on comma or is there any other ways without creating a new column?
This is not a dumb question. This one has been asked and answer many time prior.
I would do it like show on this answer

Algorithm to match IBAN to BIC

I'm trying to find a solution to convert IBANs to BICs.
In theory, all I need is to have a db with national bank codes and their BICs to compare these national codes with corresponding ones in IBANs (based on IBAN format for that country described in https://www.swift.com/sites/default/files/resources/swift_standards_ibanregistry.pdf). But the problem is, that I can only find limited and not accurate sources of national bank codes (or can't find at all, even for big ones).
Is there a way to find a BIC number for a given IBAN in another way? I also thought about first 4 letters in a BIC code, which represents financial institution - but I did not find if it's possible to convert them to "national" codes or vise versa).

What is the valid range for IMSI numbers

I want to validate International mobile subscriber identity (IMSI) numbers. What is the valid range of IMSI numbers? Please let me know about specification or any web link.
I'm not sure this question has an answer, as this number is both highly dependent on each country's independent numbering plan .
The first 5 digits of the IMSI consists of an MCC (Mobile Country Code) + MNC (Mobile Network Code) combination of which, according to this site, there are 1663 entries as of 3/24/2016. These combinations are not sequential and were designed to accommodate natural population growth. So the sequences skip around somewhat arbitrarily mainly because they mostly are.
The remaining 10 digits are the MSIN (Mobile Subscriber Identification Number), which is basically a phone number. This number is regulated by each country's numbering plan, which vary considerably from country to country.
So, there's really no cohesive rule you can use to verify IMEI integrity, other than some sort of database lookup.
Wish you luck in your search and your project!

Google Advanced News Search

I would like to Google News Search that searches for several terms and eliminates the duplicates.
Google provides several Booleans, but none quite do what I'm after.
Take :: Fiscal cliff, US debt, UK bank exposure, IMF
I want to see results for all of these in the news feed for the past 24 hours, but each searched as if it had been done individually.
Using the "Fiscal cliff" OR "US debt" etc... would do this, but would search for the exact phrase.
Using Fiscal cliff OR US debt etc... also searches for Fiscal debt and US cliff.
I want each of these to work like they would if I searched for them individually giving me all results for each term in the last 24 hours.
Possible?
Parentheses seem to do the trick. I don't see the exact result count in Google News, but I see it in the plain old Google :
"Fiscal cliff" OR "US debt"
About 82,200,000 results
(Fiscal cliff) OR (US debt)
About 358,000,000 results

Resources