1.相似度定义
我们定义，则，我们设，则,|C|=s,则相似度p=，p（0,1）
2.相似度检测算法设计
算法设计：
定义4个字符为一个字符串，将T1，T2分割成若干字符串，若剩余字符不足4个，则以空格补全。将分割后的T1T2计数，记下|T1|=n，|T2|=m，s=0；在T1中取出第一字符串，检测是否在T2中，若存在，则s+1,并删除与被检测字符串相同的字符串，循环到T2检测，直到T2中不存在被检测的字符串，循环到T1，提出下一个被检测字符串，到T2中检测；如此循环检测，直到T1中的所有字符串都被检测或者T2中所有的字符串都被删除，停止，记下此时的s；将所得的s除以n和m中最大的那个数，所得的结果为T1，T2的相似度。先以T1为被检测模板，检测，然后再以T2为被检测模板检测，得出两个相似度的数，取最小值。
用ruby实现如下：
def fill_str(str,i=4)
return str if str.size%i == 0
str<<" "*(4-str.size%i)
end
def txt_cmp(f0,f1)
str_f0,str_f1 = fill_str(File.new(f0).read),fill_str(File.new(f1).read)
a0,a1 = str_f0.scan(/.{4}/m),str_f1.scan(/.{4}/m)
n,m,s = a0.size,a1.size,0
a0.each do |txt|
if a1.include?(txt)
size = a1.size
s+=size-a1.keep_if {|item| item!=txt}.size
end
break if a1.size == 0
end
s/[n,m].max.to_f
rescue =>e
puts "error : #{e.message}\n" << e.backtrace[0..2].join("\n")
end
(puts "you must cmp 2 txt file";exit) if ARGV.size != 2
r = txt_cmp(f0=ARGV[0],f1=ARGV[1])
puts "#{f0} and #{f1} semblance is #{r*100}%"
下面是4个文件分别为1.txt 2.txt a.txt b.txt，内容如下：
1.txt
NFC East rival quarterbacks Tony Romo(notes) of the Dallas Cowboys and Eli Manning(notes) of the New York Giants now have something else in common ḂẂ they've used the same wedding planner to help them tie the knot. Todd Fiscus, the man with the plan, set
up what he called "man food" at Dallas' Arlington Hall on Saturday, when Romo married former Miss Missouri Candace Crawford. "I have a lot of football players to feed," said Fiscus, who had pizza and short ribs on the menu.
However, Romo apparently put all the tunes together. "Tony picked out every song, and when it plays, and what the keynote things are," Fiscus said.
Sounds like a very orderly occasion, but there was one wild card ḂẂ whether Cowboys owner Jerry Jones would be able to attend. With the continued lockout, owners and players are not supposed to have any contact away from the negotiating table. But Jones received
special dispensation from the NFL to attend, just as the Green Bay Packers recently were informed that they will, in fact, receive their Super Bowl rings in a June 16 ceremony no matter what the labor situation is at that time. Jones was there along with virtually
all of Romo's teammates.
It is unknown whether Jones and Romo actually discussed any labor issues at the wedding ḂẂ we're guessing this was more of a "friendly", though Jones is one of the most powerful owners on the NFL's side of things and Romo's marquee value gives him a lot of
play on the other side.
"I've gotten special permission," Jones recently told ESPN's Ed Werder. "But more than anything, (I got the) right ticket from him and his fianceẀḊ ḂẂ Romo's wife-to-be. (It's) one of prettiest invitations I've ever seen.
"So, yes, I will be there and (I'm) proud for him. He's got the best end of this deal."
Romo, who had been linked romantically before with Jessica Simpson and Carrie Underwood, proposed to Crawford last December. Crawford's brother Chace is known for his role on the TV show "Gossip Girl' and has also been linked romantically with Underwood.
According to the new Mrs. Romo, the lockout may play a part in the couple's plans for a honeymoon; usually around this time of year, her husband would be participating in minicamps and other off-season workouts.
"This lockout has been quite a dent in the honeymoon idea," she told WFAA-TV. "We'll see. We haven't really gotten there yet. We're taking a day at a time with the lockout. We (are not) even sure if we're gonna get to go (on) one."
2.txt
Officially, Memorial Day, observed on the last Monday of May (this year it's May 30), honors the war dead. Unofficially, the day honors the start of summer. (More on that in a moment.)
The upcoming three-day weekend has prompted searches on Yahoo! for "when is memorial day," "what is memorial day," and "memorial day history." The day was originally known as "Decoration Day" because the day was dedicated to the Civil War dead, when mourners
would decorate gravesites as a remembrance.
The holiday was first widely observed on May 30, 1868, when 5,000 people helped decorate the gravesites of 20,000 Union and Confederate soldiers buried at Arlington National Cemetery. (Some parts of the South still remember members of the Confederate Army with
Confederate Memorial Day.)
After World War I, the observances were widened to honor the fallen from all American wars--and in 1971, Congress declared Memorial Day a national holiday.
Towns across the country now honor military personnel with services, parades, and fireworks. A national moment of remembrance takes place at 3 p.m. At Arlington National Cemetery, headstones are graced with small American flags.
This day is not to be confused with Veterans Day, which is observed on November 11 to honor military veterans, both alive and dead.
However, confusion abounds anyway, with the weekend marking for many the kickoff of summer, and it is reserved for weekend getaways, picnics, and sales. Searches on "memorial day sales," "memorial day recipes," and "memorial day weekend" are just some of the
lookups related to the festivities.
a.txt
23l4kj23 klgjdlskgj235 3lkj 0952ru lkfj lkqejfg
2t34lktj3409t uj34gjklejeglekjfdklsafjalsfj
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
sdgakdgjsdalgjaslfjsalkfjsadlf
b.txt
23l4kj23 klgjdlskgj235 3lkj 0952ru lkfj lkqejfg
2t34lktj3409t uj34gjklejeglekjfdklsafjalsfj
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
sdgakdgjsdalgjaslfjsalkfjsadlf
测试代码如下：
ruby -EISO-8859-14 txtcmp.rb 1.txt 2.txt
1.txt and 2.txt semblance is 8.653846153846153%
ruby txtcmp.rb a.txt b.txt
a.txt and b.txt semblance is 79.54545454545455%
因为1.txt中包含非utf-8字符，默认比较会出错，遂指定外部编码比较。