From 1c03dfaebc26c1a6f871e34a1d3f31531cb4f565 Mon Sep 17 00:00:00 2001
From: sunilhariharan <sunilhariharan@gmail.com>
Date: Tue, 27 Nov 2018 09:49:55 +0000
Subject: [PATCH 1/4] Done

---
 __pycache__/__init__.cpython-36.pyc           | Bin 159 -> 158 bytes
 .../__pycache__/__init__.cpython-36.pyc       | Bin 170 -> 178 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 655 -> 642 bytes
 q01_outlier_removal/build.py                  |  11 +++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 185 -> 184 bytes
 .../test_q01_outlier_removal.cpython-36.pyc   | Bin 1835 -> 1838 bytes
 6 files changed, 11 insertions(+)

diff --git a/__pycache__/__init__.cpython-36.pyc b/__pycache__/__init__.cpython-36.pyc
index 2e5da7da778c48a9c58d7d055bbdb642fcb66418..0da7c21063d03fafbc765f0576739409bcd8f284 100644
GIT binary patch
delta 56
zcmbQwIFFITn3tDJT=;MFL=JNmZ~ct?+*JMI(!9)^jKrc$AWqEF56>^kE-pw+PSsD&
KPf4AaVhsT7xf2)w

delta 57
zcmbQoIG>Tjn3tDpxuj9lL=JORAN|nc)S_bj#H5VO;*8Yn;?ks|#N5QZ%)~tXywvje
Mw9K5;;)%)D0Qk=nh5!Hn

diff --git a/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc
index 2f9a42a105b2b26ec10c60ac4d11fa03f7624d22..895aa5939abc9d4bc72bcff29fd29c3cc21cade7 100644
GIT binary patch
delta 56
zcmZ3*xQUU&n3tDJT=;MFL=JP6c>Rq0+*JMI(!9)^jKrc$AWqEF56>^kE-pw+PSsD&
KPf48^?GFI<XA@Qc

delta 48
zcmdnQxQda(n3tF9Mtoz`L=JPAaQ%$@+*JL<oYV?^m(=3ylKcYwfTH}Y)Z~)liOK!|
DbC?i`

diff --git a/q01_outlier_removal/__pycache__/build.cpython-36.pyc b/q01_outlier_removal/__pycache__/build.cpython-36.pyc
index 8248a1626e901c868c177407620fc7ec379db180..68e8da30d99b76ac0af53976d2f42546de3df5f8 100644
GIT binary patch
delta 382
zcmYk2%SyvQ6ozLeGi?%@Y}{Dz1+p=;-4qo;Tq*8c2vLxslhUL~rn!lWk%I2rh~z1J
zA6N4zg5a4Jf*j62eE+$?oDcTOoSEmj^Ml{}>rdx>J46Uk$V<rj38X;+k5P&RdBZOl
z$m)pz(VnaZrQq!Q6<(W)1v6C9lqgU$sYvfL#lQzOJCwEeT^Jas!7U0{cY=?%k$WR~
zagocmXacI84CqnthHJZQdC??Utf{EvMzf?4@pCW)G-XmwB2rG_ooOPLktLQBBhxQb
zjo@2958+DYaagxSlC5}^K=LBIlhwN3@I@Rhq=>`r8()oGa|#~>VSg8g*hJ-bb8>WQ
gj#Uken3mn{f8v3s?Xz5pHjB?KpxX4>#Kj}}2LmBltpET3

delta 414
zcmZ9IK}*9h6vy+DcCKZ0yAJmO9D+RvUKBy}=t0nvmr<KA({;^EGT5OqaJOAV>e;*B
z$5Ze#_yIimIw%f#d4Krx%m2NQXaB_yj-n|1l<#-nufclh5kh?I3vBa-QW6eh=#z|a
zx`HP_&$c{8?`?Tzp@fo_Y{dYx8S;8UpOnD;KACtpiQ$|-#*op5?f-*#y<`n+*bN!B
zssS7>U3jIE;(V&+h1GzM9eS?)I8O>IbK(4%mW6d5*J|c`S#YsH9^@Dq@x&$$@Z38{
zKZuSn{Bs*WsHupPTrA>CVJ241;%lukky;byYAlVF=}3#Q7RE?bR9|qs7t|lsLm|-H
rR0jr|(%tHS4o+&WcWUX@QZL$)U9_%!I#qn0i!+MOv{2eKgpT(M&p~0&

diff --git a/q01_outlier_removal/build.py b/q01_outlier_removal/build.py
index ec278ba..4359f4a 100644
--- a/q01_outlier_removal/build.py
+++ b/q01_outlier_removal/build.py
@@ -1,3 +1,4 @@
+# %load q01_outlier_removal/build.py
 # Default imports
 import pandas as pd
 
@@ -6,3 +7,13 @@
 
 
 # Write your Solution here:
+def outlier_removal(data):
+    q1=loan_data['ApplicantIncome'].quantile(0.95)
+    q2=loan_data['CoapplicantIncome'].quantile(0.95)
+    q3=loan_data['LoanAmount'].quantile(0.95)
+    df =loan_data.drop(loan_data[(loan_data['ApplicantIncome']>q1)].index)
+    df1=df.drop(df[(df['CoapplicantIncome']>q2)].index)
+    df2=df1.drop(df1[(df1['LoanAmount']>q3)].index)
+    return df2
+
+
diff --git a/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc
index 5a057ffb73694628cef3ed87e03ee3a17f7410bc..9e9f97c9a14de2a7a372c553c023c14bea36a450 100644
GIT binary patch
delta 56
zcmdnVxPy_yn3tDJLilg=L=JP6RQ-(n+*JMI(!9)^jKrc$AWqEF56>^kE-pw+PSsD&
KPf4Aa5()tPBol!E

delta 57
zcmdnNxRa5?n3tDpxuj9lL=JP+H2u)x)S_bj#H5VO;*8Yn;?ks|#N5QZ%)~tXywvje
Mw9K5;;)%(j00?Lk?f?J)

diff --git a/q01_outlier_removal/tests/__pycache__/test_q01_outlier_removal.cpython-36.pyc b/q01_outlier_removal/tests/__pycache__/test_q01_outlier_removal.cpython-36.pyc
index 4c0b6c7431c6a14108ba9d55dd44059612f144a5..a8d1970dffc737fbd0405fec3901a85cb7b209bf 100644
GIT binary patch
delta 72
zcmZ3@w~mj~n3tDJLilg=$%&k+IQY_2OVUbnayD*vVN@yA&&bbB)h{m1%go70EXoAp
a#611*{G#mQg2d!h{p9?V)XfS^+^hhK6d57_

delta 69
zcmZ3-x0;XBn3tE!Sjs4B??ld3EL`cSB^!6UFshd6hZd(673(J^Wn>m-q-GbFCKV;-
ZCgx=(=IQ69mdB@M=A;&HR%YU61pp127hM1V


From 6d970374cd6c4aaec786657e489a01c12bb73612 Mon Sep 17 00:00:00 2001
From: sunilhariharan <sunilhariharan@gmail.com>
Date: Tue, 27 Nov 2018 10:11:47 +0000
Subject: [PATCH 2/4] Done

---
 .../__pycache__/__init__.cpython-36.pyc       | Bin 0 -> 180 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 0 -> 1387 bytes
 q02_data_cleaning_all/build.py                |  21 ++++++++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 0 -> 186 bytes
 .../test_q02_data_cleaning.cpython-36.pyc     | Bin 0 -> 3408 bytes
 5 files changed, 21 insertions(+)
 create mode 100644 q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all/__pycache__/build.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc

diff --git a/q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..b1a289403893eb1187e01b4bae474f475f789cd4
GIT binary patch
literal 180
zcmYL?F$w}P5Jj_KA;MM`p25ZuZACnRjfI89FeJk|n$4_9w(v+^$<|v4g4sg+@E)Jv
z{B>PVX4B8fl)MK0QkAkRq*0MG2Q#ns4-<#^uQ$P2TZYsJ@n?!spa4gW*VK_^u<97R
zHj=VfjNC*{Ni<*=^};s8ZMi~6nb3NsAfZ9>9uOr+MqG6F=Xleq>pgFW>r^mb(a|ys

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/__pycache__/build.cpython-36.pyc b/q02_data_cleaning_all/__pycache__/build.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..8f9e0adbd0b00f3e76bc928ec258029da8cf2c9c
GIT binary patch
literal 1387
zcmaJ>O^?(z7`8K+B$J)&%&w{yIG`5wGM5=Xt_X^tYLN;TkV@UEmN)j!dT>6l9n_$i
z)0E@(fArt<>?^1J3zZPOPH2_wp|<?`$@}=@{c`SPS^CqJ_dh?05PF9OZ4CGu80HTU
z9C54=>1Rx^K$~DQ5*TDSlswquzT<xFm_*&)m?YgiArqKKDlHDkLANGU(;_1oMk3lA
z^6?&%nV4<nJmT>_0Gx@L*YSi;en;ep<7a4*UVu&LsY@t5l}b>$z;13elO?snSZZn|
zE$C)dW0e$|YEe|LnR;`BGs{-0VkND$;4-(eDrr;ZO0ZJ!Wo}-(_y<ovesqE1j@%31
zCNLmG=phVq5409-u!SJ3F~{e4hX>yQd<WZrkFKIk_-P*R5fAzJJirJvk!NhQi9g|F
zj|Xf#U_BnPm?!_TJ-*CN2CT=I*>u2qe3?BMupSRs@XjX=y$wLl-v+0_RkRxov`3vL
zwCQ%V4SB{7`RqLWCESg7(JtO4Ub`J{qiwuRP9sR+@l(i8EPl&X!-a<uk^s=9KKu`g
zswwUBUSBJg3o`H0(i8N#&;>bs=u>Fv_tMx(w<P*blw4>MKVe!+!Abf^)LvCuL$YT=
zozicMT2-y!+Nbiu=X4DN+F$n<ftzYp@~WW5GAqcWFD3a#Tzr70Gg!7baFecjd>y&W
z2qkh$xovA<Tu|07Dx2a&Xcs|Us*<^|fWmUAKM|~yxdBZDUPe%j4eSY}s&YN@M^9YX
zojT3o_I0IY&Xgf394LD1@C%1q7r&rgsk;aORJ)7Ye-r>f@h8$>!hwJPx;m>0u`*35
z)fv+g1}j%TR{Eu>A)>1s41u?>?hJz`vev>F{~l^xZNRV9tDCo}H;H!7LzZjGlv<rM
zQt@Toy6h5n1Xr4CAW$&Jb9@b7AN0@hF}{NJ4PY0?Ze&`BRt4VI%$~W3)jmxw=2Dle
zfZ)`S2L?-O?#AGM-Re2))SrPMUznHhSarGdaa6QV0a)Dy(LDesT4t+4$&=;(__lm?
z^A`Qj-(lH#rbD>JC0Z~};dh`9{b$H^Rd?ohp&f{?y482Vm_~_vK9R<oImch)r0Y{W
K4N^SCN%#@P(TK|c

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/build.py b/q02_data_cleaning_all/build.py
index b56e2bc..023f1a9 100644
--- a/q02_data_cleaning_all/build.py
+++ b/q02_data_cleaning_all/build.py
@@ -1,8 +1,10 @@
+# %load q02_data_cleaning_all/build.py
 # Default Imports
 import sys, os
 sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname('__file__'))))
 import pandas as pd
 import numpy as np
+import statistics
 from sklearn.model_selection import train_test_split
 from greyatomlib.logistic_regression_project.q01_outlier_removal.build import outlier_removal
 
@@ -12,3 +14,22 @@
 
 
 # Write your solution here :
+def data_cleaning(data):
+    
+    categoricals = loan_data.select_dtypes(exclude=[np.number])
+    numericals = loan_data.select_dtypes(include=[np.number])
+    numericals['LoanAmount'].fillna(numericals['LoanAmount'].mean(),inplace=True)
+    numericals['Loan_Amount_Term'].fillna(statistics.mode(numericals['Loan_Amount_Term'].values), inplace = True)
+    numericals['Credit_History'].fillna(statistics.mode(numericals['Credit_History'].values), inplace = True)
+    categoricals['Gender'].fillna(statistics.mode(categoricals['Gender'].values), inplace = True)
+    categoricals['Married'].fillna(statistics.mode(categoricals['Married'].values), inplace = True)
+    categoricals['Dependents'].fillna(statistics.mode(categoricals['Dependents'].values), inplace = True)
+    categoricals['Self_Employed'].fillna(statistics.mode(categoricals['Self_Employed'].values), inplace = True)
+    X=loan_data.iloc[:,:-1]
+    y=loan_data.iloc[:,-1]
+    X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=9, test_size=0.25)
+    return X,y,X_train,X_test,y_train,y_test
+    
+    
+
+
diff --git a/q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..fc3c04b5e986294e8fbd449debd3d018f6e30a5b
GIT binary patch
literal 186
zcmYL@F$w}P5Jj_KA;MM`p25ZuZDsKUHWn5ZLr8|~Xf|0h*}@}vC0lPH2xbfM!+U&w
z^|y6BTg*RaRq`73OI6CgAd@1ij%ryQ9x9FVUvCF(><DBC>|33<paYy@x=ODNfr*yD
zN5_)cU?P`Dgk!)y#S57XxAg|C&Kiv;9XL4D-a{tI0*D+qBd+@UbGqxI9XvJT0V$|2
D9nUm#

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc b/q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..70ecc00d464e2e488281e492ae5170ab2bddf99d
GIT binary patch
literal 3408
zcmd6pTW{P%6o5Up*Y;j>ZPL&}O(B&QmEATx01`q)ZK)zvL82fkgXFrN*^L)pHZ$Xh
zU6rTpuizi>m*$nH`~@T=&asp1W)pIuFY(IfjK?!)=6pW0=X};|*49`4`T571rv0VO
zTo$w+!7cs(K{ZNZ%_E@oSWgVkAlPTbW>WDg&^Bo$wi4U3i+(G1lB!oF8Y6=mwI{@@
zv-+Sx9a^2}(5|q0(PyLn#qye7%Ukdky(Mp%EeuvB+F*5XW1@L$Y;opk{h3B<wEkA3
zb!JQq?<OTLw07eF_NVQ%iE@5mLOu=zgZ4s}%Q#}(=Pb$ggZMy)LD1i%feidGW<eUI
zeduWQne63p9PqvvvT%GC2Xx~sNc|yaGzw*urGB0kQyA@pV&Ai$<Lm7w2Lz0Uxct$D
zt2}~RJO{xFV4hA1)!%BKK@Dnx22|7nU7<F0KwGp*YoKjfrwz~!ZPFI#DqWz9plfsq
z++0>m=-!jr?T5HNUbw6S_hY!l76_?HG9kZ{H`+j##v4M(hBkq1z!qknd-|F-DZV5u
zVD4}<$88YfJ5PTD7?M#Jxa0{-rEoXtefL?k&r)c$O^)%wD<D)o5Q1^}bT1ELWyCCP
z>&g@??kN+g{;BaM4EsCp?q&(=iad?t-GE1M2Wj`GjQ=WzLCCscMj3$bM?yv+e78Pl
zLLmGh&jzrW?%snf|HS9>gE;O=pdt$CpBm5^j?`@&FhB3esTOQZm>@L6A>+;|!kzOb
zpTLhs3j7~HV-|l29}mtkYGdcXfKlTwk112u3-VaH5UDv6F3sFDlWr2oaM$e@%ukMJ
z8&m$g3bxuhuY=+Z6o|VN=%2cEh7*V8#0m^30T2N2>7@W1%{}%X;|9Z_@ZW_7gdGH+
z)Cb1E98@Tg7KY<Yu<1MF@21W}OuKS4WNt6xE@v|5X-UG3g!^h3q%;sNM*g=vNZ70U
z?kLM4FxlZ)ILPj%XxnPB48(1*oeCMGAyd{%ADvhFkSa6GI8&7#+ECSFbIQb}+oa?-
zhi>x>bmazpoaLA1-vgHg0?<f{=w$r)r7)hhbKQXQWpGK^;3N;Cl&``Qzky;6#X5?c
zApV2hy^GnQSv)VZ@V@vz@X*lpq(@#}m79AJx4U!1wysX>nz9;ASt$#0O2mY+UKIMg
zNV|yG=-xXM8(mLg_UP2vyaT>+7sWfA%|o-F%yGN=**r9KJ?UAnJID^xpyQW<cTrL(
zt#;<vT|Ft3c6Qoz<@kOIS<ClT&G(ax<}tP#zK<nW`GzmxCtpIbf#Md5_fdR^;v*EB
zD5e>%XkmrGv3A5Ku?WH;PQ|eu^XON+ZEtv1nTQKo%7H3Q7Vxn4^uB>&Z4}5XiKEw@
zb85KG-h+q!Txs5Uokua{SP`j8*)GaTJ4&%4mqn%HIh+O5$LhOU6z5n7LMhF&A-@eX
zIi?YgiI`(a!|^|oe+<I2B9TIUudK4RQbv%DyxR9gmEl95SD&^Y7QUW^NI74X+`q<R
VM_>YU^ad$^HP8-Oh2-BN{{rFkpb`K8

literal 0
HcmV?d00001


From 4ca8e81e443542d5c9c6d0339ff4722de9f558d4 Mon Sep 17 00:00:00 2001
From: sunilhariharan <sunilhariharan@gmail.com>
Date: Tue, 27 Nov 2018 11:54:59 +0000
Subject: [PATCH 3/4] Done

---
 .../__pycache__/__init__.cpython-36.pyc       | Bin 0 -> 182 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 0 -> 2084 bytes
 q02_data_cleaning_all_2/build.py              |  53 ++++++++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 0 -> 188 bytes
 .../q02_test_data_cleaning_2.cpython-36.pyc   | Bin 0 -> 4763 bytes
 5 files changed, 53 insertions(+)
 create mode 100644 q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc

diff --git a/q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..aa2a9a5cb7d29a5729fcd6ee4732f0d2db1825b9
GIT binary patch
literal 182
zcmYL?F$w}P5Jj_KA;MM`p25b*+KPAr8w(4IVMvDE(QFbk*}@}vC0lPH2xbfM!+U&w
z^EY)pTg*RaRq`73OI6CgAd@1ij%ryQ9x9FVUvHzk(36(Iv+FQ%0lhwjbd`vPv<a59
z3oT1#1Bu#1A{;!#6fR`4zO5Uum=TO4^z2(e=b+I5>^TFh;{7?@_Q4L0cH_Ae)E76)
BGM)ec

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc b/q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..d1b652c273e4e45f8b7a5bf9b39f89270cd946b5
GIT binary patch
literal 2084
zcma)7&5qkP5Ef-w@{i=d-K0&Ay)6pZo6U6*bg@~q0k%QWv}nK<P-&5N<WQ7CNkJUQ
zCvT2<kiHVHJ>?Y&6zEWrwf16BR07GFkE5AyD02A3^LjU@|NM628pdD7-eZIOEwofZ
zM-7U&fl)b{xCNuht*DLLNYAyn9XZ%3Ws7&BF76_Op;eEzw+Q!Hf8|n}I$IOuF6)<=
zqsQNE?BM|(;t?L>3GLACje(CSdTE^WK0{pN#R*dGHJLMhC4!hT3A#r~M*M&?B0{kQ
zU>N5a51I5Oi{f{LfB0Hk%y>+MpGZc-APZyR=c4?E(P@yrQ_i)Xe*Ww;f-r%lgTZqI
zP0N?iQfV4SW~@*F**`M2<#=8ps77F{5m=Gh#%ydV1koJ=D^=UH8k-71R3qlr*jR<N
zYC%M+wzSTT8fn)^yFu(4u^Ysx5vN8_?FYsNsr(?i^Mh#R2gb&x@`I@MW2`J#h4vEB
z&L^~K@9AC8?aW>|g}p@^r*Kvs*bT-<llEx;qq%_mH=RAyrCyD8;j3=ZDY{FO4(RZc
zx#^XNQS=(5U-V%Ijq3SI8y{$s1MR5M&;et&xYadduVm~2<3TaFhw*R^y=6Qq`3;Lv
z!*5(7s2DfMq?qh6uCyKFO4~86v>oF~(Jv;2S+o`w?EK>wczBT{JPe4)p2N0}7#?57
z<POx+8~7rMbCF^DJ0{?8VCM%SWymP*UeP=tdfQ{~CF2YKDoS{~W)zN#6T(5rDbpf*
zUy9g$#uDv9WT}7p!Q0cGe1ngElySmjw)QV1BmPghBwXDAf01({q%IoZ)*|o26!&hu
zVDG>M9$e~Unfc#`X%@>h9+ii&O2+@0$>^l7Opz!neJeA43=Qlpz&?{iq>ETa%1aq%
zLFUtJov>8dB9G=wDl?&~zhr>tQ53>Ni^qYogIK_7D2qZOCr8S8r9Fj0*&whqQ_gw^
zfB*m<a5CqN`>d)ucm)2Lgg?U5$`Vh&AzQ`=m{&)wO@;@m{}|r^&Y;v)G^-E%MJNc@
znZn@L9Pw~JwB|NGjI1HI@l%rD*C6cRKtF>*yNe3ebx^kn3fP~9{1Vpu?>}a*Ve`$>
zT!j2Jks&l9X1~VrO`4E^&2&YEGP{Jo7zR*%my)HaF13k_SMYacZ%@zsgKF&)&i%95
zJP$dYCTlf3oIM-sx_b?sfhM=FY42pSMq48^LOt{lUO>zNDqr(nv?ad+e<#PvPKcl+
zl|xWf8<sU$%Rb2XLHY)66Dg)}TLASyXCap5zEhWw)S6^*#KZaYKXRSko9nd9RJL@)
zi~sNGv47`EtNrM!Np;{5QeWR;-SUdlODOflrO+#d*76uc`9OCMb@zqt^a_;!bgz2^
dnChHIG0i!9swt)u=sKu_T+@ZtLoUe2*1sNyQq=$e

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/build.py b/q02_data_cleaning_all_2/build.py
index e20ff7b..c28b53b 100644
--- a/q02_data_cleaning_all_2/build.py
+++ b/q02_data_cleaning_all_2/build.py
@@ -1,6 +1,9 @@
+# %load q02_data_cleaning_all_2/build.py
 # Default Imports
 import pandas as pd
 import numpy as np
+from sklearn.preprocessing import LabelEncoder
+
 from greyatomlib.logistic_regression_project.q02_data_cleaning_all.build import data_cleaning
 from greyatomlib.logistic_regression_project.q01_outlier_removal.build import outlier_removal
 
@@ -11,3 +14,53 @@
 
 
 # Write your solution here :
+def data_cleaning_2(X_train,X_test,y_train,y_test):
+    X_train['ApplicantIncome']=np.sqrt(X_train['ApplicantIncome'])
+    X_test['ApplicantIncome']=np.sqrt(X_test['ApplicantIncome'])
+    X_train['CoapplicantIncome']=np.sqrt(X_train['CoapplicantIncome'])
+    X_test['CoapplicantIncome']=np.sqrt(X_test['CoapplicantIncome'])
+    X_train['LoanAmount']=np.sqrt(X_train['LoanAmount'])
+    X_test['LoanAmount']=np.sqrt(X_test['LoanAmount'])
+    
+    lablel_encoder = LabelEncoder()
+    X_train['Gender'] = lablel_encoder.fit_transform(X_train['Gender'])
+    X_train['Married'] = lablel_encoder.fit_transform(X_train['Married'])
+    X_train['Education'] = lablel_encoder.fit_transform(X_train['Education'])
+    X_train['Self_Employed'] = lablel_encoder.fit_transform(X_train['Self_Employed'])
+    
+    X_test['Gender'] = lablel_encoder.fit_transform(X_test['Gender'])
+    X_test['Married'] = lablel_encoder.fit_transform(X_test['Married'])
+    X_test['Education'] = lablel_encoder.fit_transform(X_test['Education'])
+    X_test['Self_Employed'] = lablel_encoder.fit_transform(X_test['Self_Employed'])
+    
+    
+    numericals_train = X_train.select_dtypes(include=[np.number])
+    categoricals_train = X_train.select_dtypes(exclude=[np.number])
+    dummies_train=pd.get_dummies(categoricals_train)
+    dummies_train_1=dummies_train.loc[:,'Dependents_0':'Dependents_3+']
+    dummies_train_2=dummies_train.loc[:,'Property_Area_Rural':'Property_Area_Urban']
+    dummies_train_final=pd.concat([dummies_train_1,dummies_train_2],axis=1)
+    final_X_train=pd.concat([X_train, dummies_train_final], axis = 1)
+    
+    final_X_train=final_X_train.drop('Dependents',axis=1)
+    final_X_train=final_X_train.drop('Property_Area',axis=1)
+    final_X_train=final_X_train.drop('Credit_History',axis=1)
+    final_X_train=final_X_train.drop('Loan_Amount_Term',axis=1)
+    
+    numericals_test = X_test.select_dtypes(include=[np.number])
+    categoricals_test = X_test.select_dtypes(exclude=[np.number])
+    dummies_test=pd.get_dummies(categoricals_test)
+    dummies_test_1=dummies_test.loc[:,'Dependents_0':'Dependents_3+']
+    dummies_test_2=dummies_test.loc[:,'Property_Area_Rural':'Property_Area_Urban']
+    dummies_test_final=pd.concat([dummies_test_1,dummies_test_2],axis=1)
+    final_X_test=pd.concat([X_test, dummies_test_final], axis = 1)
+    
+    final_X_test=final_X_test.drop('Dependents',axis=1)
+    final_X_test=final_X_test.drop('Property_Area',axis=1)
+    final_X_test=final_X_test.drop('Credit_History',axis=1)
+    final_X_test=final_X_test.drop('Loan_Amount_Term',axis=1)
+    
+    
+    return final_X_train,final_X_test,y_train,y_test
+
+
diff --git a/q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..8e52c374ab79c7547f538e2f49253f9a8c4ef620
GIT binary patch
literal 188
zcmYL@F$w}P5Jj_KA;MM`p25bEwH5IMHWn5ZLr8|*(QFbk*}@}vC0lPH2xbfM!+U&w
z^*41rTg*RaRq`73OI6CgAd@1ij%ryQ9x9FVUvGvk^aL_^cAZXK&>l`9T_tKmV1gxZ
zp=HTzFj1RGgoDSJ!i7wR+qyxkvqs}cd-g4A=dgiHk_8aib4Fan`*XVOgB=|0#tTwV
EUxzp~1^@s6

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc b/q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..a7f6bf06208e553ac702d558a21ffea31dda16dc
GIT binary patch
literal 4763
zcmd5=TaVL56!tBSbKOh!!a@O;7HBG(kfj$&p<)FTAysKrT7g_7V=}W@r;Z(F9MnYg
zDZKO_^r`RtH+`PhKIJb|s?>ArB<pOm=_c?%ET1_Uk3Y|Rb1pM;tx(8rm;U+j_ktw-
zB_)13=(pkKe}W(qkv++hp;SDjty-!q5tV43)=pU|&?S=c^mf`xN4oAA?TnR)^t6|4
z=d7G8QMr{T#z3|Tw9qP&49N}@&@;3U)upj~qFXE0s<mdVTP16Qrsx`7rzN^E(4J_M
z+BK5fky@Jr$=af8BtO!(>H0)J1^T$n&V7j#Nb!|Kic}q_)@dR?l*%hZ7?o5n%fk4c
z^6-ws=}>_hs8}IR=-3U9I=<^S?OmuUOjTi<pc{HFWj3Sj;DzHsRk2CK{jTRZtjRmH
z(Z7K2RK39Q?GB^FZG>*%+g(2brliu~FRb*}`1rxyp$wn4@^@*6`h@x+w`*4ETNZRE
z3w!n+qmKOpd*=8J?DY7D-MtMr_dw8Sax8_&M0q7yDozfR8c7iy$`nZx14<ofhcZoa
zBoCz#Q)6Yw3R#8!S+Yjfq0EsI*?=-nHpv#01+q;}L0Kd_K;CIlLelO|^VGmW;14@y
zz<md9z6&ChLU|znBEOVcN~pe+iM%5Xprg>q#Q%L|OB$eEN$?4+&Jb>>f#{#V_cJgW
zQepzR-8K?rUL;q{`|b<sgI3m98rm^zEOHL#K-<0NUB?ruNBy!QG)}#Jq2bj0A+Z%`
z_IIy(5VUEPcYW78aF`3X<5wRA>?huF8noI72!+{ex;%6nFwIRyIma37u%HE<sXo8D
zYfqVl?RcKOTMdC39z8|vNR!=4rzg(MdSFk6R)X$v9EzkGa({OQ)0O!EQxkze(L|gB
zB{>oCG@wIINLTAa6<YN_pYToSzSH$W6HL5Ixk&=k4??r;gpC8U8L>azqO2ay1j_+d
zSz!=ZBvwR$#K&S4G;wz3Fh+oU2b#n*$UsBw#59cGpRj;-guyIg!2|`&5DYR@T53yc
zrHCBrXr2k<Q|kS14k-XDH^W|sn)?AW84bJ4kNGfNZa(ceK5@8-zVH=u+Vtravln#1
zQ0b^G#1Qi$FQ>(7Y?u%D1D}WB%T(x(ZKOab9U`<wz^F*=BZkOK;Dm;sF3T~CM+m$;
zlfbx(Hoj-aA2<(#=174g7iC56e{o_m|BLmu(aJVpSRxGrWsd8!E%?W_QJg}7M#N5o
zIEKLwPQoApzYiEm3U4k;A%eZ79QKZKD0GOVF6Bah7?rQb<Osg}MwlFdy`4<z&^z$!
zgfXHAG84rdWBh-2j&UnT7+lVsW2`9xC%FPUIgXN$s3BE1Az4rIgjRPSK~x>?L9_n#
z?~)<05p*#-lK`^CVXC0x{2k>HGj0UA0yBirJjk&k6Zx+lt)&(Zb`frDz8@nyxCD`w
z2y_XYWtrJTW(M&2*Jg%Y1Qd~-5K~(wh$Ti;`wZ~L)GQppl0ZvxGkgRc7~+cX4RK>c
z;Tw8(3^SMJh4;~bz)^1`c)77?U_XRemKBDqV9t6&tYpTl9A_|d1I^)hYPS-+EF5GK
zIJ1et#2p8jmO`{dcuCcVzo8T1nz4U%D!lLm2Nq&;qYJD!ufdB9<G?7yuy;Tgsp#cn
zETufpQlb}eErrSZHeijN=+>gfGKI7_Ex*FIqk%0&3%dUlEt^>J35a7znY%ig3<Vr0
zT3U}{?!tj32~;ny2*b90SOnX)$l7*0AYBjjylvx-E`Gw+u;;i@VrNjCMe#0*_fT9$
z@gWLKl?($8`xM1>6o-rGNW-lN%Yayio8zj{kPRKKRCMWvI=-S-<ujHZulb|&E)3W}
zhG;wt$1hUPx~@l}ov3vO_KZCz4BDRitTJaSS(&+&ti(tuK(+Qxj#RUc98HXHaQQFN
zu8(IEVYErJ3^-RHHl9OdqGJl&&BBQR96Jd4v5<R_+px$;aP!D6U{~_7U&kq&UBY@?
n(Xo$E)KDOovq%=k&3HDy1m~7GLK<>@d?{Jb40%J|n$-LYsFoz`

literal 0
HcmV?d00001


From 06e68db3cdecf9f37a6794e9a4710f0f5e95b690 Mon Sep 17 00:00:00 2001
From: sunilhariharan <sunilhariharan@gmail.com>
Date: Tue, 27 Nov 2018 12:16:11 +0000
Subject: [PATCH 4/4] Done

---
 .../__pycache__/__init__.cpython-36.pyc          | Bin 0 -> 182 bytes
 .../__pycache__/build.cpython-36.pyc             | Bin 0 -> 1279 bytes
 q03_logistic_regression/build.py                 |  13 +++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc    | Bin 0 -> 188 bytes
 .../test_q03_logistic_regression.cpython-36.pyc  | Bin 0 -> 2302 bytes
 5 files changed, 13 insertions(+)
 create mode 100644 q03_logistic_regression/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q03_logistic_regression/__pycache__/build.cpython-36.pyc
 create mode 100644 q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc

diff --git a/q03_logistic_regression/__pycache__/__init__.cpython-36.pyc b/q03_logistic_regression/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..d3dd25b121f007eee3924ec9737a7862e2d2f61a
GIT binary patch
literal 182
zcmXr!<>e9={u|8z1dl-k3@`#24nSPY0whuxf*CX!{Z=v*frJsnuVnp<{M=Oi;?lg#
zoQ%YxOdw9o(+|%t$}TQQOitBL&QD3z&&f~EEH24RjxS10FG?*g&dkq?FDS~-N=+`&
nFElWY$0DyEAD@|*SrQ+wS5SG2!zMRBr8Fni4rF;T5HkP(PR27j

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/__pycache__/build.cpython-36.pyc b/q03_logistic_regression/__pycache__/build.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..c76a3513d72458db92fe9c87f795e5fd7a83911f
GIT binary patch
literal 1279
zcmbtTOK%e~5MJ*)Nt3kcyGL%lB&8KMgg}J@sHzY>P?<`LS-a`B-bZbxB4{fig)2XV
z8^6a_PW=lYB*rF@(y9j}HuCJ3Z@lA~Z^nD=cJuS2-|v3;j`PErdqTiZVae}cs6!EV
zFoL(s-Q2?-GMvZ#Jivj)J`eK<M-~S>&KtO4ambr_3%4we_(I;sZHr^x$-B6V9EQ#p
zX=8%$5?ea&(I#z8T)-{1WVyC^pM-cBui#a@hS%{1>#|k0#@5-!#6R(8EM1`OFAm<M
z=*Zdb+(6dO;WpBpBSi{I1U*U#X9AesK{?E%%F;J%D43L4SpXZq&ML;Uf&pGk%i?ro
z97#@;$UXwsEk}xHOeBKk<rU#TwJA{~NjW1$Rt(LaTST(^>46~zyd*_Z2}ZM2nV6%(
ze#2-#l~+1CFz%PnZxCdhLL2PU0Bi6*EcptIf)-HsE4ua-nmE^ia;aAbr!Mtt|I{72
zGyq>%hv#06YH#8q2e4nG3A&C{P<wSWaj)Y#264i}8xtN9UJ#&7nI}>a#dIjGM0%M|
z1%g`V6bhtbmD~=62OkU97?Yv)PBW#WX?-FvksCTXPLv>7p+f*n!iMpT000E;-94og
zqHW^E)wB~j99!A6o95pfu?fzf?+0gP&IWQ+Wc-YX3>GN{Z%c6@E0VH7T2cmr|84MG
z{t&1<XQ>)o?(8M^_y-?G8K?bfEIM$^3M{z@#z9RKq8N410-z?c-t8?NR%WQAb}Nck
z<pqp@DEc4)oRXT6Duz}dyfX8d$s!`jAw6yd#EUc6tX!)m?#}3(!Q`dV3aXz$AI3zL
zInO@y?@51FUjK6EVRC0u`?fS;F0|hN_c*)B?P(^QPa{Pp(u@C!l<ZDp#U@rP7@>)|
z3=Ml%%)#2HqO3GJ)@UqxuuCi(V>%;NjWK<)GcjkjfYb3~P<h1J6EmmMP%(6G?&mh)
KUyN4Jn*Rrgyn3<#

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/build.py b/q03_logistic_regression/build.py
index cdbd506..5b797a9 100644
--- a/q03_logistic_regression/build.py
+++ b/q03_logistic_regression/build.py
@@ -1,7 +1,9 @@
+# %load q03_logistic_regression/build.py
 # Default Imports
 import pandas as pd
 from sklearn.preprocessing import StandardScaler
 from sklearn.linear_model import LogisticRegression
+from sklearn.pipeline import Pipeline
 from sklearn.metrics import confusion_matrix
 from greyatomlib.logistic_regression_project.q01_outlier_removal.build import outlier_removal
 from greyatomlib.logistic_regression_project.q02_data_cleaning_all.build import data_cleaning
@@ -15,4 +17,15 @@
 
 
 # Write your solution code here:
+def logistic_regression(X_train,X_test,y_train,y_test):
+    logistic_regressor = LogisticRegression(random_state=9)
+    scaler = StandardScaler()
+    pipeline = Pipeline(steps=[('scaler', scaler),
+                           ('logistic_regression', logistic_regressor)])
+    pipeline.fit(X_test, y_test)
+    y_pred = pipeline.predict(X_test)
+    cm=confusion_matrix(y_test,y_pred)
+    return cm
+
+
 
diff --git a/q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc b/q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..0273a442229709d1f5b3c4fc7611b7de31fa1d1c
GIT binary patch
literal 188
zcmXr!<>e9={u|8z1dl-k3@`#24nSPY0whuxf*CX!{Z=v*frJsnuT1@n{M=Oi;?lg#
zoQ%YxOdw9o(+|%t$}TQQOitBL&QD3z&&f~EEH24RjxS10FG?*g&dkq?FDS~-N=+`&
tFElWY$0Dy^l3H9+tREkrnU`4-AFo$Xd5gm)H$SB`C)Ez*gkm6O005!eHKG6j

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc b/q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..0681446bfaadc7845a0ab2875a79cc83f66a2985
GIT binary patch
literal 2302
zcmbtVQE%He5Edm_Rvg<&n>1O5VbcQSK|o-)=~IVoXpsd)fFLg$u!1Y9vLfTiqa=r<
zg4$@G+&{CwvDf_rz3wT0VZeahNr_{(3A%0)csw1cBk#M9$A_IxYjf@IU;gkc>rZR$
zfp6<EH2pgWZgC<lN~+$bHuSc1QkS{}#~j&6J?d4uE1RiL{Yr1hrL;v`pnJS2m(w<F
zS9V`^(k|^1OOW9TUz!rSDprSgc#AJjZO~g{wX%8GKlkWex<=ROJ-R_Rg)i2{J+UD+
zr|xriW^G?v>1R(Z-sYWm7Vijm>d-Atezd~w1;k+;hotm>658xDEkN&ZW>^$U!LlS9
zfMI395bdEUFN{otij+w6b0#lra0K=>nGX_ek{AXCN@$(r8JN2RVNMDunHuPkh$q|V
zu`hF$MI$A65*wHo6<M{Y;JsL%)8-KlAO3Jbz;%opy?Su1$I$dH2vMZ~;3wREXHkbc
z+=brd4emkT;7#sB@9`xFxK(aqphq8WGR9T#LFYcSeQ0_Qgs}{nlDFiIHMEWMhH$cF
zO<^%uXzu&O-ms<)W)hZhj}^MRa6wGII(Q8s8^HsJzDPx8bg;vBgQv;4$UqBSh3ly%
zh;o@}EtEMpE0`=DDYDQmT`lBE=^}Q&S!xxg{q?qgnx~?#i!70+OeN4**8e$Izv>Z-
zML*8D0C+!01)!Y|VKe=+`wycV#`^})(UpvDnbI4L%Wa%?)Bn*d#JjLfjqzKKPbS~o
zf_m@cjusMw4y(j`3Cet8>a^eyoF&0IzHngH<k1X&E>2h>O#pl?gbsKfWVs1aX5!Oe
zP-SCrM(A8;Mzz6LXsa%CY6S%*q84{RE5E)qluM_26Q<RW5P;;-Cqr`OXNCxJ3`<gl
z$QPghMS-lw9y**D7il|3O8a<nbPWZZzrc(~A~?xapoA$@R^t*RI(V5C>1f=`I8%y^
zU+xCuya4u!%ZX4Gf*l<;%GH{!hx#znhGnrR?KmwP)!(#v`FrW&Iw7f{dk?knR@CY}
zMOEoR2N<8mXf4trHkmy6Wc=RuxPaq~SsODPhxQao{34@X!06<`48{fEcn+jOM%7%J
z9|I;Ke-t_|os_-41&*DM!BJ6n8@y<&Qx=(g_g`VCJROJ4(vPByr6P*TRurW<FC^;i
zD8d?4k0?CNY8}NE3e3eE|ErwhAy6$4sA#NcKJgpydX+W&oO&Y$l+dL;;tJ0~=|lN4
z7_q`&=ssK+V`lPHCa-!o6sN1=)jPYtSM%cPj*6I+y;nsdxxyn)5B`^@Xz$V&a)Lhj
z_r4xP@7;=*ei|5TquyUt<(o-{*PyA&3-y&?Ji;q)sk#x7@^F8td{mNWB^lG^v&blx
wWYhx%;G)g(l>rpQ4|!F=d{b6R|JypP-y_T#lLx6G?YXxFx=+^02I<=W08#Lm$N&HU

literal 0
HcmV?d00001