numpy - Python - Data structure of csr_matrix -


i studying tfidf. have used tfidf_vectorizer.fit_transform. return csr_matrix, can not understand structure of result.

  • data input:

documents = ( "the sky blue", "the sun bright", "the sun in sky bright", "we can see shining sun, bright sun" )

  • statement:
tfidf_vectorizer = tfidfvectorizer() tfidf_matrix = tfidf_vectorizer.fit_transform(documents) print(tfidf_matrix)
  • the result:

(0, 9) 0.34399327143
(0, 7) 0.519713848879
(0, 4) 0.420753151645
(0, 0) 0.659191117868
(1, 9) 0.426858009784
(1, 4) 0.522108621994
(1, 8) 0.522108621994
(1, 1) 0.522108621994
(2, 9) 0.526261040111
(2, 7) 0.397544332095
(2, 4) 0.32184639876
(2, 8) 0.32184639876
(2, 1) 0.32184639876
(2, 3) 0.504234576856
(3, 9) 0.390963088213
(3, 8) 0.47820398015
(3, 1) 0.239101990075
(3, 10) 0.374599471224
(3, 2) 0.374599471224
(3, 5) 0.374599471224
(3, 6) 0.374599471224

tfidf_matrix csr_matrix. find on this, there no structure same result: scipy.sparse.csr_matrix

what structure of value (0, 9) 0.34399327143 ?

without vectorize can recreate matrix, more or less, sequence of operations:

in [703]: documents = ( "the sky blue", "the sun bright", "the sun in sky bright", "we can see shining sun bright sun" )

get list of lists of words (all lower case):

in [704]: alist = [l.lower().split() l in documents]

get sorted list of words (unique):

in [705]: aset = set() in [706]: [aset.update(l) l in alist] out[706]: [none, none, none, none] in [707]: unq = sorted(list(aset)) in [708]: unq out[708]:  ['blue',  'bright',  'can',  'in',  'is',  'see',  'shining',  'sky',  'sun',  'the',  'we']

go through alist , collect word counts. rows sentence number, cols unique word index

in [709]: rows, cols, data = [],[],[] in [710]: i,row in enumerate(alist):      ...:     c in row:      ...:         rows.append(i)      ...:         cols.append(unq.index(c))      ...:         data.append(1)      ...:

make sparse matrix data:

in [711]: m = sparse.csr_matrix((data,(rows,cols))) in [712]: m out[712]:  <4x11 sparse matrix of type '<class 'numpy.int32'>'     21 stored elements in compressed sparse row format> in [713]: print(m)   (0, 0)    1   (0, 4)    1   (0, 7)    1   (0, 9)    1   (1, 1)    1   ....   (3, 9)    2   (3, 10)   1 in [714]: m.a        # viewed 2d array out[714]:  array([[1, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0],        [0, 1, 0, 0, 1, 0, 0, 0, 1, 1, 0],        [0, 1, 0, 1, 1, 0, 0, 1, 1, 2, 0],        [0, 1, 1, 0, 0, 1, 1, 0, 2, 2, 1]], dtype=int32)

since using sklearn, can reproduce matrix with:

in [717]: sklearn import feature_extraction in [718]: tf = feature_extraction.text.tfidfvectorizer() in [719]: tfm = tf.fit_transform(documents) in [720]: tfm out[720]:  <4x11 sparse matrix of type '<class 'numpy.float64'>'     21 stored elements in compressed sparse row format> in [721]: print(tfm)   (0, 9)    0.34399327143   (0, 7)    0.519713848879   (0, 4)    0.420753151645   ....   (3, 5)    0.374599471224   (3, 6)    0.374599471224 in [722]: tfm.a out[722]:  array([[ 0.65919112,  0.        ,  0.        ,  0.        ,  0.42075315,          0.        ,  0.        ,  0.51971385,  0.        ,  0.34399327,          0.        ],....        [ 0.        ,  0.23910199,  0.37459947,  0.        ,  0.        ,          0.37459947,  0.37459947,  0.        ,  0.47820398,  0.39096309,          0.37459947]])

the actual data stored 3 attribute arrays:

in [723]: tfm.indices out[723]:  array([ 9,  7,  4,  0,  9,  4,  8,  1,  9,  7,  4,  8,  1,  3,  9,  8,  1,        10,  2,  5,  6], dtype=int32) in [724]: tfm.data out[724]:  array([ 0.34399327,  0.51971385,  0.42075315,  0.65919112,  0.42685801,        ...         0.37459947]) in [725]: tfm.indptr out[725]: array([ 0,  4,  8, 14, 21], dtype=int32)

the indices values individual rows tell words occur in sentence:

in [726]: np.array(unq)[m[0,].indices] out[726]:  array(['blue', 'is', 'sky', 'the'],       dtype='<u7') in [727]: np.array(unq)[m[3,].indices] out[727]:  array(['bright', 'can', 'see', 'shining', 'sun', 'the', 'we'],       dtype='<u7')

Comments

Post a Comment

Popular posts from this blog

android - InAppBilling registering BroadcastReceiver in AndroidManifest -

i'm implementing inapp billing in android app. works well, however, i'm trying decouple broadcast receiver activity manifest. suggestion in android's trivialdrive sample: // important: dynamically register broadcast messages updated purchases. // register receiver here instead of <receiver> in manifest // because call getpurchases() @ startup, therefore can ignore // broadcasts sent while app isn't running. // note: registering listener in activity bad idea, done here // because sample. regardless, receiver must registered after // iabhelper setup, before first call getpurchases(). currently there's class extends broadcastreceiver : public class iabbroadcastreceiver extends broadcastreceiver { /** * intent action receiver should filter for. */ public static final string action = "com.android.vending.billing.purchases_updated"; private final iabbroadcastlistener mlistener; public iabbroadcastreceiver(iabbroadcastlistener listener) { m...
Read more

python Tkinter Capturing keyboard events save as one single string -

is possible store keyboard events 1 string? code below stores 1 char , prints it. card reader or bar code reader, contains collection of character/ string not 1 character @ time. goal save char pressed text variable. from tkinter import * root = tk() def key(event): text= event.char text+= event.char print ("pressed", text) def callback(event): frame.focus_set() print ("clicked at", event.x, event.y) frame = frame(root, width=100, height=100) frame.bind("<key>", key) frame.bind("<button-1>", callback) frame.pack() root.mainloop() currently, creating text variable , everytime key() function gets called, , text stores last character typed. you can define text module level variable , , use module level text inside key function - text = '' def key(event): global text text+= event.char print("pressed", text)
Read more

sql server - Why does Linq-to-SQL add unnecessary COUNT()? -

here tables: create table customer ( cid int primary key, name varchar(32) ) create table ord ( oid int primary key, cid int foreign key references customer, address varchar(20) ) here's linq-to-sql statement: var aaa = c in db.customer select new { c, o = c.ord.tolist() }; and here's inexplicable query generated linq-to-sql: select [t0].[cid] [cid], [t0].[name] [name], [t1].[oid] [oid], [t1].[cid] [cid2], [t1].[address] [address], (select count(*) [dbo].[ord] [t2] [t2].[cid] = [t0].[cid]) [value] [dbo].[customer] [t0] left outer join [dbo].[ord] [t1] on [t1].[cid] = [t0].[cid] order [t0].[cid], [t1].[oid] i want understand how rid of count(*) part. it's entirely uncalled for! linq needs count(*) in order determine number of entries returned o = c.ord.tolist() .
Read more