From 1ce969de16a7b9e00fd51f52f786f6f6d290f50b Mon Sep 17 00:00:00 2001
From: clydemendonca <clyde.mendonca01@gmail.com>
Date: Wed, 12 Dec 2018 16:09:36 +0000
Subject: [PATCH 1/3] Done

---
 __pycache__/__init__.cpython-36.pyc           | Bin 159 -> 157 bytes
 .../__pycache__/__init__.cpython-36.pyc       | Bin 170 -> 177 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 655 -> 635 bytes
 q01_outlier_removal/build.py                  |  15 +++++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 185 -> 183 bytes
 .../test_q01_outlier_removal.cpython-36.pyc   | Bin 1835 -> 1837 bytes
 6 files changed, 15 insertions(+)

diff --git a/__pycache__/__init__.cpython-36.pyc b/__pycache__/__init__.cpython-36.pyc
index 2e5da7da778c48a9c58d7d055bbdb642fcb66418..3a7c63a114f5fc0e9fafc46b07e83778aedddf29 100644
GIT binary patch
delta 55
zcmbQwIG2&bn3tF9fT3W_L=JOhFa3=C+*JMKoXV8c+|<03{Ji8u{qX#v?BasNWT1F{
JO6tTEYXH#%5^ewh

delta 57
zcmbQsIG>Tjn3tDpxuj9lL=JORAN|nc)S_bj#H5VO;*8Yn;?ks|#N5QZ%)~tXywvje
Mw9K5;;)zMt0Qi&>ga7~l

diff --git a/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc
index 2f9a42a105b2b26ec10c60ac4d11fa03f7624d22..55880042c21f857088b8ec425e12833d073e77a5 100644
GIT binary patch
delta 55
zcmZ3*xRH^=n3tF9fT3W_L=JQ1IQ@+L+*JMKoXV8c+|<03{Ji8u{qX#v?BasNWT1F{
JO6tUDe*oHl6086K

delta 48
zcmdnUxQda(n3tF9Mtoz`L=JPAaQ%$@+*JL<oYV?^m(=3ylKcYwfTH}Y)Z~)liAnwd
Db7l~S

diff --git a/q01_outlier_removal/__pycache__/build.cpython-36.pyc b/q01_outlier_removal/__pycache__/build.cpython-36.pyc
index 8248a1626e901c868c177407620fc7ec379db180..a89dce20314c6a8177264ec26fbd4f5c895b6023 100644
GIT binary patch
delta 393
zcmZ9H&q@O^5XO`2cDr3|X*~#i1NUOV-bAdWC&infur8F4gj#nu**~pC*y6#vMfw!p
zeF9&>TVA3kFM4sZ2#N{$nBUCDWM)<SvK#$o(|bMWgwb>3N!SRX5xn9Oq(mHd(JkWS
z5ic>2Rgn!ufNB*vKvwzVEpcQb@RcM&Cng40X_>f-mcSRtudfgB{=+0(y?yM3n?c~6
zD=LSHn#<XwU!Bds$J)*2RL-Wc@NI2#I-@~58>VSIWsp9VOeG@N$wrF)%WY-9{a#S+
zz#^a5y&JIiSbWb#A|zL`!)WhH<x!DR1_i~%`W>GwJhOWX1mTvsgpgg_Cb<h4-_cG=
zC8vcZDc5c;C?B)pPB&s$U~EQ&L35Q>?&{+|Qil%OJxUax$Ku!mD$Hmu@$k0&4PH%Y
AEC2ui

delta 410
zcmYk2u};G<5QcqrQktrzU8BOr(uE92OcfG>kpT%NmMX+naht@&cBBqL9lC?6GPCnC
zOiX?V9sn3PcOg3amwtWs-&y_%Uc%_4*Xw-Bx4W-R^t=iQAs2YfD@si`Od%#S!s!~G
z06lAXiauy~W}$|f)@;oHvoqv=LQHDlVNB)$PGUIcpD|>#$&MNj+iTW@p52g9Yx8@&
za-A!cmgj}4N~;0iItpC-u}VuTbK$~;mZfz8*J|NHS#q&N9_1LB@YJRb@Z3Ad-ihvE
z`1>|_P=!d+Tr87IVdhpXl54GIBC{sR)l?cQvxyc{EsT+>+<w66F!KL2Kp`;LRTmC+
nrTg1MIvi`?p!dA=z4%QH*+=Wz<3jN&7iScmX`!@f2fD!@ePLsE

diff --git a/q01_outlier_removal/build.py b/q01_outlier_removal/build.py
index ec278ba..76c0aed 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 +7,17 @@
 
 
 # Write your Solution here:
+def outlier_removal(data):
+    qApplicantIncome = data['ApplicantIncome'].quantile(0.95)
+    qCoapplicantIncome = data['CoapplicantIncome'].quantile(0.98)
+    qLoanAmount = data['LoanAmount'].quantile(0.975)
+
+    data = data[
+        (data['ApplicantIncome'] < qApplicantIncome) & 
+        (data['CoapplicantIncome'] < qCoapplicantIncome) & 
+        (data['LoanAmount'] < qLoanAmount)
+    ]
+    return data
+
+
+
diff --git a/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc
index 5a057ffb73694628cef3ed87e03ee3a17f7410bc..2d672c6d153f032a7f5a57446f9c3cf2c817b232 100644
GIT binary patch
delta 55
zcmdnVxSf&1n3tF9prK&QL=JQ16#b0++*JMKoXV8c+|<03{Ji8u{qX#v?BasNWT1F{
JO6tUvPypZ{64w9#

delta 57
zcmdnaxRa5?n3tDpxuj9lL=JP+H2u)x)S_bj#H5VO;*8Yn;?ks|#N5QZ%)~tXywvje
Mw9K5;;)zM200=D;>;M1&

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..4302c1b9afe5a99b2a0a7e60cd790ff79c888159 100644
GIT binary patch
delta 71
zcmZ3@x0a97n3tF9prK&Q$%&k+IQY_2OVUbnayD*vVN@>B&&bbB)lbf;Oi9g6%}dG8
aOHR}e&o9a@E=WuUisz@KZdPF8W(5Fbuo!p%

delta 69
zcmZ3>x0;XBn3tE!Sjs4B??ld3EL`cSB^!6UFshd6hZd(673(J^Wn>m-q-GbFCKV;-
ZCgx=(=IQ69mdB@M=A;&HR$}621pp0E7h3=T


From 688b9b52797ca9e5d48a449e4179c6e2f64e3d28 Mon Sep 17 00:00:00 2001
From: clydemendonca <clyde.mendonca01@gmail.com>
Date: Sat, 22 Dec 2018 14:58:00 +0000
Subject: [PATCH 2/3] Done

---
 .../__pycache__/__init__.cpython-36.pyc       | Bin 0 -> 179 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 0 -> 1282 bytes
 q02_data_cleaning_all/build.py                |  27 +++++++++++++++++-
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 0 -> 185 bytes
 .../test_q02_data_cleaning.cpython-36.pyc     | Bin 0 -> 3407 bytes
 5 files changed, 26 insertions(+), 1 deletion(-)
 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..3a728e2caf520f3349321485cfb2270b6a0437c2
GIT binary patch
literal 179
zcmYL?JqiLb5QVc~A;K0O!^Zg2ig*GW3mah!nPH8aWMUGFJ(9QZ9Jb!V%A^n<d>@b3
zymi}7p3CV;l)S5cX-e4@(x?c<LCl){L&Q4&^*PH6JCfQCNRB+R*Ge8a_9-Y`c1Dtx
zJCm}}5UGoll3@=a@|AR!x5WxjWd+)j^2T?dtd#)9n+z}y!)v^8jCV`h`kV^l3wV1m
AV*mgE

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..befbf8609a485a79518e4c4c409146609d92b28a
GIT binary patch
literal 1282
zcmaJ=&5qqP6t<n*{M~zJG8D>GLbE7ca5wiBkXRvp#0aSpst{eYidtE2>`9Y4KVv&p
zZ<VaVb~d~PZ^0w5_Lc>jS760)E=8j*8q3GW$9B&5eaHTCy-xoA`RwUe3xxhcH$Ir7
z@1X17z;MK|MA>1)85XD_l_xM*|EA?+!iR<rdly+SpAEAJMo-3dk|pz+FPHTyTh0DL
zuIr=h2qO_x$2^#DmWs4G;USME0{%**F2-Y?{ED&<IDUjS%WJ@e9@>b~i&6?oH`t~|
zv9h5?XhU@;O9Qdvwl}gAiYigJub6~T_@M53Bh;I_oEf&0Eo*3}1TS+_whir@Tng3*
zzRmS38~)^$ADmxfNS8a~z9sYk`VY|cA7BPlp$Uz6Kn5EEGjDPucH=3U@EtS~<C{el
zaLft!uDsDRVHFKt6%U}9+sXYaGUC^t4WR8gdUiJY0~!1~s7l<q|NP~ICW6uBFOLzN
zUk!L51>uW{j1I$^2eZEz5Dy?{V2*~x5I7}yeC0p&$8e0scuZV$2#08hhvXsvJ(Gvl
zuYtC56gQqds!OfQrl6o%FLV}Z5Z8!eZ+-T<)K12{m3`f46~c*ZIqUSuFw^S-0`UF$
z?R5;phGg-fB;^}nSBf>ft*M5cVsmUyXdw)3`dX+mr>-E@m!&q=)50jHvvOC_lBTm>
z!x~FEZsT*<_oHI9$krF7p#ZJX7p<yos6;1OE`UtSg$?qy$(gY}S8b;}hh_;Amt}P`
zXO_Y)QM7=8$r(i!mWNy7{^599{KVqD4WGDdWn%+y;G)@Z|26;)1b#>AF68;=+3rPK
zi(M}FT+{+M1CZV0R=w06;M?T@53XJTRDrh=g%Y4QT)$4WmB`KR^8Oby5IWyRR|zKN
z?s;EIzU}sQ?MVE;mO@5-8w~Oz{0T|%ZL-2=Bn2ChQ>^a6%H|emy4RMp+WH+cFKocN
z4v^W9m#P6VZO}n+m{zx7yIp`l-Cmu5pM9)f!kbXd)~PC~77`!=+F4TjJ%Hac)7G+l
zzI_j;?dAQ?={xUnJ2O(g%kt6NEbtmZhTJLA6o>jpN-$2{i;U+7>*%XfI5l&dCBI=l
ebp)cX0c|hEH?HpbLog}6ha+e!*GK|3^8N*QS$6IK

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/build.py b/q02_data_cleaning_all/build.py
index b56e2bc..b565b4a 100644
--- a/q02_data_cleaning_all/build.py
+++ b/q02_data_cleaning_all/build.py
@@ -1,3 +1,4 @@
+# %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__'))))
@@ -5,10 +6,34 @@
 import numpy as np
 from sklearn.model_selection import train_test_split
 from greyatomlib.logistic_regression_project.q01_outlier_removal.build import outlier_removal
+from sklearn.preprocessing import Imputer
 
 loan_data = pd.read_csv('data/loan_prediction_uncleaned.csv')
 loan_data = loan_data.drop('Loan_ID', 1)
 loan_data = outlier_removal(loan_data)
 
-
 # Write your solution here :
+def data_cleaning(data):
+    
+    
+    
+    numeric_data = data._get_numeric_data()
+    categorical_data = data[list(set(data.columns) - set(numeric_data.columns))]
+    
+    imputer = Imputer(missing_values=np.nan,strategy='mean',axis=0)
+    numeric_data = pd.DataFrame(imputer.fit_transform(numeric_data), columns=numeric_data.columns)
+    
+    for column in categorical_data.columns:
+        categorical_data[column].replace(np.nan, categorical_data[column].mode())
+    
+    data = pd.concat([numeric_data, categorical_data], 1)
+    
+    X = data.drop('Loan_Status', 1)
+    y = data['Loan_Status']
+    
+    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=9)
+    return X, y, X_train, X_test, y_train, y_test
+#     return categorical_data
+
+
+
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..da6cfcbee9528a7837b9c271136a5e2f9131115c
GIT binary patch
literal 185
zcmYL?JqiLb5QVc~A;K0O!^ZKa74ZZ%7B<2Vl5vfjWY$b9_DJ5sbJ%(dE3<|8;QM&I
z=B?{`@?1_&s^mTBm#UOqK}JPZ9n`GaKU5m#zdmQUa3C<gw-hK^jz+_gQ!AY|WoJv^
zxp9)6!9-0W5yug`#8)yIZi^LKoi!RyI=a}P_8u}x7C>~-WyE>czs8%u_U>sr{7XT7
E0l;)K%K!iX

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..37b6501b177e96b53a250f36c2323ca45b6ad862
GIT binary patch
literal 3407
zcmd6p&2HO95P-QPMN+^1iQ~jg+OTOIKt(FW?x8JE6e*H4Fc7#v)1tv50K;%s3RQ}f
zcb7(Gq$lMm`UZWAKF3~r@+<Vz8Pc|7+4WCzF$K;nm%Fn&A7?f5MYCD^Yh&r<&o#sN
z+nBi=Xg`Eo{sDp+j3fpnpv}Y_TGS%gXC?NqLMza=StW6XE_I83C-H_=S|tW2gBo)u
zgw}a|&|n^`PE2T5c)jRzQNMI(leXvrU8GBNnJ)~kO^m_H;QGX%t9)_h>Bduo)mZ(V
z!Rp+aSo9_%&y99tANFT#w~6+C;!-^dC5QGxma8P@A`pC-?S{#|34@@&$wCzdQNqJC
zPW#Z&>T}i0lOz;<IpWdYT^!I&vM>!sg0nbMah3*oT1?@r6UklbKEu~rkM{`}jd1zH
z3s-pvw|oYI7r>~=2s7Uq)M6I1K?5r0fUYo?d7vFuWi`+)tFs1Zk2P5fbd@czMbI_2
z1a2<tC3Nrc?DivEA1_?igZmNOaub9y6q%6U$y;M!D(fv_WX+hsHed@g&pmV1m=s?U
z7O=MkniDRFy*p2S0~m@kAGqX0o+|0Dv-|$jc$cTpYTE+iLsUTMdMG6q>d8(XCfZ7P
z+BUT<dD7E1QvK5qO&Io%>b}W_yc;EB#=#=X(kSfyl8ImCD2$*#V;s2mW2xc@K3iXK
zDUtn1WCPescjv)oaO~>^VUlzeK#_$EP7LUb#`-o6n4fp(Qw#PbZ4id#k-g3dz@4)u
zAES>(3i=;GW0ro&9?#7YX={7mf>C>4A3<ik7v_obAy9KJeU|xYru<>3qBnlOKz@8g
z+nVARRj}1IMIDrApg`KCK>zfuQ=B+3C$GVP5&!`JpI->T;oKwtF>Ejz3jcj*K-59}
zDRW>A>_LSQ<zO_<#G1Li_ruhA2x(u9N8Ima!WUfSA}vYyvGiY!!jy&5$H4zyghT%7
zzCX@#h)aGj7S6H1F59kNEaPxXZlzL%X~ebjB0%T0Ibzz5GQoAFhc<Ne$egxu={70(
z&7oUA1zowp0A~fI`47Nli2w}JA|~1U@<JF-+PP{##WJ|0U2sx_aVl2eNnA&<isA-}
zn;`y!-M#bKp;<gHv+%z9Kk(4d)ucyWT$Y=E9=E%5#5ONa?257)Pg!XPQcBFFcAgjd
ztVlbL*!bRi6B}PmV($3F*}Mb33LnLLoXrEXpUrW*{MkG(bT#QYusg^O)1VU<f_Gk0
zD6Mwp*j+v;ly<h;b?pT~3Rx=%bS(&m8OsxFH-Z2wuJVmoz)!w}VhzPD6d$7a7{w<j
z)=^9|T+zY`L15{KPjV52N4$#XdiLS3c-vm1PML@cTH4FgSQYTF^rYWGu{I7>HcaBz
zoil2<&d!64V6HOnyw2l<39N{8rEC{fr5mSMkgKB7p`PF&3$XaE7PUDRf>23|Y$R^O
zOo3@cU?LV+(g^&I6rX~iPApR>@3m9bR@w^FF|GYrR2c#E(fYK#QTTcqBJF)sa{m^K
V9f=9hGaID*)j)e>1(JV@{0r3|pK1UA

literal 0
HcmV?d00001


From f5ffb8142009ed9074ee8f41fcaa197e5b7d1ba1 Mon Sep 17 00:00:00 2001
From: clydemendonca <clyde.mendonca01@gmail.com>
Date: Sun, 20 Jan 2019 08:25:34 +0000
Subject: [PATCH 3/3] Done

---
 .../__pycache__/__init__.cpython-36.pyc       | Bin 0 -> 181 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 0 -> 1278 bytes
 q02_data_cleaning_all_2/build.py              |  31 ++++++++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 0 -> 187 bytes
 .../q02_test_data_cleaning_2.cpython-36.pyc   | Bin 0 -> 4762 bytes
 .../__pycache__/__init__.cpython-36.pyc       | Bin 0 -> 181 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 0 -> 1316 bytes
 q03_logistic_regression/build.py              |  19 +++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 0 -> 187 bytes
 ...est_q03_logistic_regression.cpython-36.pyc | Bin 0 -> 2301 bytes
 10 files changed, 50 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
 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/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..4720e43b9d6135dc426166106982b89082e05984
GIT binary patch
literal 181
zcmYL?F$w}P5Jj_KA;K0O!^Y^^ig*GW3majGnPH74nV7_4kK`>phpo4;GAYCl@A3J~
zU$yPzxttzF$-C;8rj%_Vjfzn0#jM#qM6B~)pOZYZBdP6x<j5m?t>l5@H3g;1&PdX7
zZ&EfIBK46{GVCElzL3uHw&(y=R-i2@Z+s8RTIeKz@g@Vz!|)of9pl~7rhcb__yXM&
BG4=oe

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..829cc54c8a55100f025e629c3df33dea31049c69
GIT binary patch
literal 1278
zcma)5&2HQ_5GJYLmG{r7?b?9bLr?_R9(G&z8Wbtw90~-v<Y0qHA<!bdkwQ^2$ql@P
z_N1Qj1bKj7`YL?{UVF-=50F!bN*lOEQB(qFIK%n*W;pz2GRc1Z^87^-BlH`(^MQ;W
z!7^`vaKy1h6c3!x0639h9nlE;vp~jmLX%-1%CsKQ5k>-68IO8Q_rzW`<}put0{fBJ
z8+3{1FJd~OQ#zw_x=-(kPsFS`=rP1e`RFG^55=LFiMiMp_xcdD_&Hi+?*I>aw!m)u
zlwAq=L=_De8srIQmgR*MOqFU4idnO<vJ^VkqHbO@`Qv?Wv62l_d8-943tKiS->AVy
z@MU3MyX2`q{{H+OhBQU!lfjq4@*Q~$%lr<+q8{zS+n+ns;War&KfK*VT~uMv#$9BQ
zB_IVo1|EBUjCy>6cFFC1FN6K@F74tjJwa7yqblyPmmZ><M;GY!L6;0Rr@Pc95OLIv
zs<cbIHP`_MJAQ%esLH@ASrd+XvWx5<<Q)6{bzwE>p!XQv%r8LV1af=sG2jm|!FxKy
zB|Yx(j(n5>Ey#hdtRr1;9YWWGXWEyqeNz@Q7s}FDx1X<t&DEwBx-4?vTNfB%T~ai1
zQ!C>l*0yr%h-zJEZnV_|-09@Hw7Jzxnd?T^j<npxMWYI4UDWDQ0W&XID_jijLc!^C
zNFdB2bIGN5ER~C4BLG6O{YwEGP|*0gR7_sFaDX$u3y#hqg;+NLmdOurA+Ja3h!npd
zeG2H_=c|`ZEmnoxa#2He4{UYO=vSs?1?XYeV1m|AKh&(nS_@<R6t%jkL}6DqU!4uH
zKMrEfq|DD&SDRAuWxI8=zl~{rwhvj&7eEk+@f4@{l#KB)tQa4WInMBuq@a!Q9M6gV
z9PAcHE^e9P%;+hIE`loCtscXk9-3G10JK`dfCHq$qf_gl@$NBz-7?$Mvb<XUM-9t=
ztzkJ-paY-ryZ;~Q>--<D@Xt8@Ww}%f#&iGdM#HNdE-mD`X&t_Fc&iV<N*{W1<jJR=
r%sug48=5iP+8y#49Jush-SCYR-+HvhyZRQ6{ZBHWDLx=ske>Vj^+0H$

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/build.py b/q02_data_cleaning_all_2/build.py
index e20ff7b..31c95e4 100644
--- a/q02_data_cleaning_all_2/build.py
+++ b/q02_data_cleaning_all_2/build.py
@@ -1,6 +1,8 @@
+# %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
 
@@ -9,5 +11,34 @@
 loan_data = outlier_removal(loan_data)
 X, y, X_train, X_test, y_train, y_test = data_cleaning(loan_data)
 
+print(X_train.shape)
 
 # Write your solution here :
+def data_cleaning_2(X_train, X_test, y_train, y_test):
+    
+    finalX = list()
+    for data in [X_train, X_test]:
+        numeric_data = data._get_numeric_data()
+        categorical_data = data[list(set(data.columns) - set(numeric_data.columns))]
+        
+        for column in numeric_data.columns:
+            numeric_data[column] = numeric_data[column].apply(np.sqrt)
+            
+        for column in categorical_data.columns:
+            le = LabelEncoder()
+            categorical_data[column] = le.fit_transform(categorical_data[column])
+        
+        data = pd.concat([numeric_data, categorical_data], 1)
+        print(data.shape)
+        finalX.append(data)
+    
+    X_train = finalX[0]
+    X_test = finalX[1]
+    le = LabelEncoder()
+    y_train = pd.Series(le.fit_transform(y_train))
+    y_test = pd.Series(le.fit_transform(y_test))
+    
+    return X_train, X_test, y_train, y_test
+data_cleaning_2(X_train, 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..911aedb6c3403626d817b2929774e02bc7d1380c
GIT binary patch
literal 187
zcmYL@F$w}P5Jj_KA;K0O!^Y9I74ZZ%7B<2Vl3_QRWY$b9_DJ5sbJ%(dE3<|8;XOXT
z`kT6*J=gQIDtV9kr7C4#kV%nMN42O950%FGuTKXZ2Lj^<OM#;0Xf&KSU8UEi>}&}<
zw@$J%n5a!8;y7ZT_(CSbZPlRFS)=i!ql+zS@3DbQk_8Z5bQy8k53lJqu)TZQjSr-t
Ez9^bBUH||9

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..7227768d7d54db9a36985a9cb040a2940b7c8fb5
GIT binary patch
literal 4762
zcmd5=TaVL56!tBNbKT9}V4;9Z3p5qYlBE|)p<)FTAysKrT7fJiV=}YBsbhy32a$+A
zg_r(=KJ~r-qCcU}^V+BUg+BG1aguem*>n?lAePUZjK`m6zB!kfxmqmd{;ZWB|6Wv-
zzm>!<4f-v(`ClN2LR4R|RVX!IYwNbIDnuuSZ?sc(3Uq~}{B%2GXQZC?&34w#N<HJ}
z+Ic&#DpYM1h&fR0A}zK`BujDw4fHH6%DN1ekJI*wy=s^3HGAFOpeb6WYjmA%42;Ld
zq_#}*+e&M5px9fqObR3Y1YMixCqW;#*}kWcA}PIANQvqL-9AOs2TEmS2%}Q!6;+r&
zQXbuQIUQ<H0~IU8ja;YUQ#bH}rn3uG#i=T66Lurtqs(En9lmgVs46vSwBPl8mo<5Z
zHu~rBogF`P1E<3%@fwjA22MASz?4)Q{DqzQ79ZcgGgRTzcHs`~(16e&;!e$8|DJ^%
z%A%gL$EfT4#GblA13NwbVRvuA&3zD5PL8b+m1wUOTgS<P(jY04hB8Gm#Dp@9v_qL8
zc~XGVjH$7+WQDB4{~RfkH7N6Booql^Ae&?h$|5;IPC{8C+d$qav5utOndYg1gTNnl
z%!K<k+<X^Aq(tgK{Y8DHw6sWnr4n^p89+y&lZpTP+LkgvyHem2Mx7zt&;ZdtclRYQ
z8c|{ax!pDrWL+Sat$W@J8h}<YSO(fLY%KCF=Rn)t=Uvwqx=(|OCJavfePQ6#{GqTF
zX!g&`oo8X2?lk-!p@2xjpyBR34B5}T<2Il^BorpE>G8;Gz$`Zz<s2ul!@?HyW#{>o
zU1!QF9M|`q-JJ-C;qobJhnnnGJ3VoB)&plUvUTVk$B`(ysrGkg5M7-QFf|VdB<JBQ
zD9L$<X94|r!ns}_>d>nH<%DiR_uZ}^SzzH^$}JLFK^R$WH)=eynv(qK78U(qB3K@<
zDjI{hqOcMQ<UJOvpoudxhcN==+t4JYK?NFW$EIQY{)7cIBaCH|1q&3gLa@h3Yw0bc
zl_F}CM(a%2o>uRFcR&FcxfS&~)Y=c3#c0%JLClBcaqCIP4T#GvbcJu2+on%0TfMLg
zc1lNWA%s{LcqJoNW4pZ1?*}{rSEeHU$UzE()*-@Zgp7*RK4OUM1Wp+E>53Y&c!<DD
zGYO2l=-_)!{DE^YXpR&pYDv}9{#VB)^S@Z{8m(*th9xpEQ0968+k$`W1d5X=(1_S6
z5Jxch(Qz0=;12;KN#TuUDMYY$l*8U(4n-OQsYkg;Kak~{F*$;-ycH%#VDBcAY3LpJ
zb;20Y1DS<ljxqkfJIA=yLkuqG&N0>$fs<T;og7C=$kULin~<v~dBUiB4<V`!_n=w-
z+7HQ)*a*9roJk1T;viMfar%~d#EKh1uD}c-3?Fi=$V&fp<Wg$!U?*{7^ZgXr!4-(I
zM4(IHEX&L$GBbeBzcDlH0-%W8gqYegK`b$%+82N~re@&)mIPXoo8d#~zz`RNZ;2aI
zhHvQE5zJhi7v4t$0!O``;N|+Ff&BnxSymXbf;sCgv63CLa+JZ$4KxSisohNQvT%?|
z;LIil6L%bBS{gAD;U!fc{)SG3YsUVSsqn&|xv&tM8(m<<c@<t{7zainhP?;6NXeIv
zu$0O?OOY?)S_+f*Ex;N((alATWeRC=T7HFZ%YiLL3%dUlEt^>J8HgiDnY%ig3<VrW
zF0Dr~cj3U21gck7gy}c|EP@?J<Q%6RlCF<>!Etay7e8TT>^W|f*l84JP`r=g0~D7~
ze2fB9CBuNjK1Xp4#la$4YPc0)SrBV*b6ho=s+oo>C9kxpk1yG(a@tPEYkryDh1m_f
z2#tr~_(kez*Yk<oh}ySd&)9RLu<d(Kt8=!J)tOt#YK)WuRD19ENHyos$;1c;m;WN;
z1$Z<O$xWJV!m$Ez@E9U1PbqLW3kL>p>LAodLhVVnVbVx&^T^I)R|>IT$0?j$#ClxO
mu}@IcP#~AHNEYUecs9QV=T<mEnrdNuX*tkLbwk~n)cgxXQzS$H

literal 0
HcmV?d00001

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..6306291c9fdb8039b0bc38c61cdd96839c800e9f
GIT binary patch
literal 181
zcmXr!<>fkHC>X;41dl-k3@`#24nSPY0whuxf*CX!{Z=v*frJsnuO$79{M=Oi<ebWs
z)ZEm(l>EHpME&smqU_>=#AKj&eoCr-PJVi3aY<%!d{Jt8QEG8<W`16LK~a8IYI2Ew
mp@DHc7J2>n_{_Y_lK6PNg34PQHo5sJr8%i~AiIl!m;nIv4>FPf

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..785b689e071fc658d7371d55db8cef5642baf78a
GIT binary patch
literal 1316
zcmbtUO>Y}F5GD6hlBLM8W!G+Rz3C#k&aE(l7<LW@3?J%448p~NS<>n%<Z?|?K@><Q
z<=TJIOa2nCIr%T-)FExPPT)g<uE5!+M{+pxX1FhUy$(A#y;dngU(ll`1N;hx`3np~
z3=2dsydx3Gn8es~u}EY}(ts0@$(-Z?r=lg>q#bZ3I&y<-1e}Ym?2#Tu98aII)(Vq8
z?@u>bn{`$Z;5P3EZo|K?GxD5lk!>;{J7kyl_%<K#9lpCtu9G#(Hdyx)B0n&EiH5y9
zh!kB6v3quDY0ao+mn9Wk1G9CmCY7;O`5T{TZcL?WU^h!u-^~1<LQ<=%55V=+%!-QZ
zLUXC!Qvp<uQA>+ba9UUO#IN~=C=NgV?1`~Zv@ROWSyfu^b5@5voQ+EJ-sR_h`t0-$
zLzHFWV}+*;!{ZYerUAn+i&#8~m&u1;7s)C5{cxFDv_i{lkxlUeEz$|zN2`F2P~h-C
z9H1j~yZHv$h$Srj6EE{cK8;~lw2CkSoGkDPFIzTU#EaG{0@<<)*FSC=QI)i|XLYG0
zch650{f_GR@Q<aM)%MN@4Kj2!sH|jREVX=?=?G}oyQyqpVaD7jEp_a<7&w<dD5HJ4
z!?w$>{IaTc831lz!F){s0D?ey#7x$vu&OY=k}kVvNFR<21-|FCOQm95xNIKu|BUoL
z9P+go{{h*IOEG6$a)`Aq>G+M(?@U8W;9(CG?Y=Rf-s{c)A8N`=JH9=5Q9Lm}emkoK
z8#Qyc{lx4xoNTthpr{jdaE1qX4^S6|{t4LOp352^y>U^)$c}jjjiu`m$OGq=zB%hj
z(2@++bE2wYO%aj|l7a#c+a!m^Elp6>9YYFpYLyh#+tCxTtuq_l9{f^#*XW~=m4@#C
zawq@Y&tdWHWuQVlc=BHm_xgW%Dh?lf<=$6IbIJ;Tm)hZ~`GZ};v}zo_a(J%2>e}nC
q2j1*><0TK>@W}iKoNFC{oHM~+`W7{NU|P8Q?+;4X#Vsh_Zt^$Xs*2eF

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/build.py b/q03_logistic_regression/build.py
index cdbd506..0433830 100644
--- a/q03_logistic_regression/build.py
+++ b/q03_logistic_regression/build.py
@@ -1,3 +1,4 @@
+# %load q03_logistic_regression/build.py
 # Default Imports
 import pandas as pd
 from sklearn.preprocessing import StandardScaler
@@ -15,4 +16,22 @@
 
 
 # Write your solution code here:
+def logistic_regression(X_train, X_test, y_train, y_test):
+    
+    columns_to_scale = ['ApplicantIncome', 'CoapplicantIncome', 'LoanAmount']
+    
+    for column in columns_to_scale:
+        standard_scaler = StandardScaler()
+        X_train[column] = standard_scaler.fit_transform(X_train[[column]])
+        X_test[column] = standard_scaler.fit_transform(X_test[[column]])
+    
+    model = LogisticRegression(random_state=9)
+
+    model.fit(X_train, y_train)
+    
+    y_pred = model.predict(X_test)
+    
+    return confusion_matrix(y_test, y_pred)
+
+
 
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..b718393fcad35df54bbaba050dd6a10d496dd627
GIT binary patch
literal 187
zcmXr!<>fkHC>X;41dl-k3@`#24nSPY0whuxf*CX!{Z=v*frJsnuMGW+{M=Oi<ebWs
z)ZEm(l>EHpME&smqU_>=#AKj&eoCr-PJVi3aY<%!d{Jt8QEG8<W`16LK~a8IYI2Ew
sp@DHc7J2=W)Z&t2{rLFIyv&mLc)fzkTO2mI`6;D2sdgY26az5>07E@A`Tzg`

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..3e1344c279d6d88c3e98a914dec23ff6d11b5b69
GIT binary patch
literal 2301
zcmbtVL2uhO6c#1fRvg=L+B8{)VN(O-ARw@t^weP+T4X^HAjo9{RuI8dR%Bdxl;n_9
zP#f9F{WJR)!>;=Wy6%*}u+!d?630#xblqf#e14=p@_p~)<I{G#`TOqb$)9!0`pcSo
z;J@`8Ui}9&+~P!9l$5<qZRl<3BrbIcjybZH)M>r6UD-%H>Xmj)E+tLc1iQ`~aye<y
zR_XU-JL%94u>={e@TDoCUC|w`@+M!N+F&<DxAfJqKdI9P^dViNkLWtx5S~~QkHosz
zn7SwK%-g#0($8L6yv5t^E#4OH)S;W4{9py0ONhfd4oKnsD74vUT7cc=%&;($f~9df
z0LRLLBiut*mKzxh6)KTr=S*JO00jPrG8@F&#1RY(l+ZfPQgC+$!kp$(GBwa65lyxb
zu`e^0h9f0-92uAw=4rX8;Jrwn)5Z}FAAWyH06M~rUO%|jb9nU*G@?ua&`-Gi-l7h7
zxC_0@YrGD9jW@UleVs2sz|CR<13mh1lM$|h4}yE}?!&A1pfQ#qQ}Qdhu!gpAE(j-^
z))W?lh35XB+UwTT!A!z3?y*9+OBb5SR|oGPWFxo_(dUUsjrO<sj{h<~7b#eQt8hKl
zfTmbxS_@?k&T=LTM~XDC3s(zyTDZvFZ<gx9w7)O<Z?i=7BRS>*5P6nHtp8J{e$gWq
zfj{E{*!>_AKz25Sz4XtXJPq#{?;9XTmu+~@l-_7uY~i%Kz7J;+K7d_n4Bv7*GWq%*
z(t95VTF46mmU;OSjQPA&S-}%H%Yk!z>A<YXvl;zdoU&XRALN<~?eok}Gvg=BL~s2;
znT*95fpe1>)dH-*RvqZn3O1OBs#yiAczkatSD<<Wrd5;>pyb&nQ*w<nV+19J6{%$8
z3ot;TpjBfJ9Zrmks-4rMeLOk3p#qMdZ^k3xpJvKe!sIHgc=2QHzfSXHH14IGDaFRG
zcl>ddgLuW&L?{aWwhn4Vw-W22K1{V?X(S3eN{U+fHEmpdUbwhUKq~5Pu@>CRTD7OJ
ztUL&S@o9|KB28kG$%{|s?>6EBk27Vh&v+c#Q>gHZiuxHwC%ZEm7l`8-hzb=|X=#28
zl!W?G2wpe|dv^~V+aIH&l<pRwXe?6}nSA?SX(*A7gGS+nVagH_hD9?BlZ@vQ?N%6K
z393dEo@TX%%_cUOi#h$5ImJVun$V!3v7mXxtHG~cdV|kteZ)W#y0AxF;aMoWJdF)T
ztS}h55BJ5GnJkg<o8BG8>AHCJ&YtX5y12fgLMCPJO&&|G@W|7H|AiFpU16an=!<`k
zwHw~L6t8e<7;LTDUs>cEaf-K~Da#A>lwdr>8*iz+5K;1QeW`3zkXHp6)5fdNC>E!*
u4hGOg8{=yS7|0*$s-*d@tQ6iiRa(D8nl&a5QbSsEzb4ooStIMDWB&sqnwTH}

literal 0
HcmV?d00001