In [ ]:
plt.figure(figsize=(5,5))
sns.heatmap(df_train_dummied.corr())
Out[ ]:
<matplotlib.axes._subplots.AxesSubplot at 0x7f9b594bc278>
A train accident data was obtained from internet and the data contains many variables like people Survived data, sex, age, class travelled & other data. An attempt has been made to implement Logistic Linear classficiation and find survival probabilities.
Import Required Libraries
#data libraries
import numpy as np
import pandas as pd
import matplotlib
import matplotlib.pyplot as plt
from pandas.plotting import scatter_matrix
%matplotlib inline
import seaborn as sns
sns.set(style="white",color_codes=True)
sns.set(font_scale=1.5)
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix
from sklearn.metrics import classification_report
from sklearn.metrics import accuracy_score
from sklearn.metrics import precision_score
from sklearn.metrics import recall_score
from sklearn.metrics import f1_score
from sklearn.metrics import r2_score, mean_squared_error
from sklearn import metrics
from math import sqrt
%matplotlib inline
The Total Number of Columns & Rows of Data imported. Totally 619 x 10 data points available.
df_train = pd.read_csv("/trainaccidata.csv")
df_train.shape
(619, 10)
df_train.describe(include="all")
| PassengerId | Survived | Class | Sex | Age | Cost | Stations | AA | BB | CC | |
|---|---|---|---|---|---|---|---|---|---|---|
| count | 619.000000 | 619.000000 | 619.000000 | 619 | 619.000000 | 619.000000 | 618 | 0.0 | 0.0 | 0.0 |
| unique | NaN | NaN | NaN | 2 | NaN | NaN | 3 | NaN | NaN | NaN |
| top | NaN | NaN | NaN | male | NaN | NaN | A | NaN | NaN | NaN |
| freq | NaN | NaN | NaN | 388 | NaN | NaN | 443 | NaN | NaN | NaN |
| mean | 310.000000 | 0.394184 | 2.310178 | NaN | 34.940226 | 2192.970921 | NaN | NaN | NaN | NaN |
| std | 178.834188 | 0.489070 | 0.834574 | NaN | 17.663780 | 3169.612906 | NaN | NaN | NaN | NaN |
| min | 1.000000 | 0.000000 | 1.000000 | NaN | 0.750000 | 0.000000 | NaN | NaN | NaN | NaN |
| 25% | 155.500000 | 0.000000 | 2.000000 | NaN | 22.000000 | 547.000000 | NaN | NaN | NaN | NaN |
| 50% | 310.000000 | 0.000000 | 3.000000 | NaN | 32.000000 | 1035.000000 | NaN | NaN | NaN | NaN |
| 75% | 464.500000 | 1.000000 | 3.000000 | NaN | 47.000000 | 2111.500000 | NaN | NaN | NaN | NaN |
| max | 619.000000 | 1.000000 | 3.000000 | NaN | 71.000000 | 35351.000000 | NaN | NaN | NaN | NaN |
Now check for null in the data
df_train.isnull().sum()
PassengerId 0 Survived 0 Class 0 Sex 0 Age 0 Cost 0 Stations 1 AA 619 BB 619 CC 619 dtype: int64
Remove the unneccessary data
df_train=df_train.drop(['PassengerId'],axis=1)
df_train.head()
| Survived | Class | Sex | Age | Cost | Stations | |
|---|---|---|---|---|---|---|
| 0 | 1 | 3 | female | 0.75 | 1329 | C |
| 1 | 1 | 2 | male | 0.83 | 2001 | A |
| 2 | 1 | 1 | male | 0.92 | 10457 | A |
| 3 | 0 | 3 | male | 1.00 | 2738 | A |
| 4 | 1 | 3 | female | 1.00 | 768 | A |
Check Null again
df_train.isnull().sum()
PassengerId 0 Survived 0 Class 0 Sex 0 Age 0 Cost 0 Stations 1 dtype: int64
There is null in Station onbaord., it can be dropped or substituted. Am dropping it.
df_train.dropna(inplace=True)
df_train.isnull().sum()
PassengerId 0 Survived 0 Class 0 Sex 0 Age 0 Cost 0 Stations 0 dtype: int64
Check the datatypes in the csv
df_train.dtypes
PassengerId int64 Survived int64 Class int64 Sex object Age float64 Cost int64 Stations object dtype: object
Convert sex into categorical
df_train_dummied =pd.get_dummies(df_train,columns=["Sex"])
df_train_dummied =pd.get_dummies(df_train_dummied,columns=["Stations"])
df_train_dummied.dtypes
PassengerId int64 Survived int64 Class int64 Age float64 Cost int64 Sex_female uint8 Sex_male uint8 Stations_A uint8 Stations_B uint8 Stations_C uint8 dtype: object
Converting into Indicative variable
df_train_dummied.head()
| PassengerId | Survived | Class | Age | Cost | Sex_female | Sex_male | Stations_A | Stations_B | Stations_C | |
|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 470 | 1 | 3 | 0.75 | 1329 | 1 | 0 | 0 | 0 | 1 |
| 1 | 79 | 1 | 2 | 0.83 | 2001 | 0 | 1 | 1 | 0 | 0 |
| 2 | 306 | 1 | 1 | 0.92 | 10457 | 0 | 1 | 1 | 0 | 0 |
| 3 | 165 | 0 | 3 | 1.00 | 2738 | 0 | 1 | 1 | 0 | 0 |
| 4 | 173 | 1 | 3 | 1.00 | 768 | 1 | 0 | 1 | 0 | 0 |
The total number of male vs female in the coach
df_train.Sex.value_counts()
male 388 female 231 Name: Sex, dtype: int64
Station Boarded on Details
df_train.Stations.value_counts()
A 443 C 119 B 56 Name: Stations, dtype: int64
Let do some heat map
plt.figure(figsize=(5,5))
sns.heatmap(df_train_dummied.corr())
<matplotlib.axes._subplots.AxesSubplot at 0x7f9b594bc278>
PassengerID doesnt have any role, better to remove it
df_train_dummied=df_train_dummied.drop(['PassengerId'],axis=1)
df_train_dummied.head()
| Survived | Class | Age | Cost | Sex_female | Sex_male | Stations_A | Stations_B | Stations_C | |
|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 3 | 0.75 | 1329 | 1 | 0 | 0 | 0 | 1 |
| 1 | 1 | 2 | 0.83 | 2001 | 0 | 1 | 1 | 0 | 0 |
| 2 | 1 | 1 | 0.92 | 10457 | 0 | 1 | 1 | 0 | 0 |
| 3 | 0 | 3 | 1.00 | 2738 | 0 | 1 | 1 | 0 | 0 |
| 4 | 1 | 3 | 1.00 | 768 | 1 | 0 | 1 | 0 | 0 |
Now again, performing heatmap
plt.figure(figsize=(5,5))
sns.heatmap(df_train_dummied.corr())
<matplotlib.axes._subplots.AxesSubplot at 0x7f9b56179c18>
There is a strong correlation between Sex and Survived stats; Now lets built a classifier, using Independent x & dependent y variable.
used_features = ["Class","Age","Sex_female","Sex_male","Stations_A","Stations_B","Stations_C"]
x = df_train_dummied[used_features].values
y = df_train_dummied["Survived"]
Now, its time to split dataset into trained and test
x_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.3,random_state=1)
Checking the shape of test and train datasets
print(x_train.shape)
print(x_test.shape)
print(y_train.shape)
print(y_test.shape)
(432, 7) (186, 7) (432,) (186,)
Applying Logistic regression
LogReg = LogisticRegression()
On x train & ytrain
LogReg.fit(x_train,y_train)
LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,
intercept_scaling=1, l1_ratio=None, max_iter=100,
multi_class='auto', n_jobs=None, penalty='l2',
random_state=None, solver='lbfgs', tol=0.0001, verbose=0,
warm_start=False)
Storing the Prediction in y_pred variables.
y_pred = LogReg.predict(x_test)
Computing Confusion Matrixs to Evaluate the accuracy of Classification used
metrics.confusion_matrix(y_test,y_pred)
array([[95, 12],
[24, 55]])
Lets Check more accuracy
metrics.accuracy_score(y_test,y_pred)
0.8064516129032258
The computed matrixs has the accuracy of 81%. That is for 100 Passengers, around 81 passengers status are accurately known as survived or not survided.
len(x_test)
186
Lets Plot other parameter from Confusion matrixs
print(classification_report(y_test,y_pred))
precision recall f1-score support
0 0.80 0.89 0.84 107
1 0.82 0.70 0.75 79
accuracy 0.81 186
macro avg 0.81 0.79 0.80 186
weighted avg 0.81 0.81 0.80 186
Algorithm has decent precision & decent recall scores. The Coefficients & projected lines can be generated.
LogReg.coef_
array([[-0.93651253, -0.02470958, 1.2435027 , -1.24361931, -0.56096349,
0.40195066, 0.15889623]])
LogReg.intercept_
array([2.97946363])
Now the Intercept is sensible value and we can check the Probabilities.
df_train_dummied[used_features].columns
Index(['Class', 'Age', 'Sex_female', 'Sex_male', 'Stations_A', 'Stations_B',
'Stations_C'],
dtype='object')
Now, the first column will be probability of entry has not survived label and second column entry that survived label.
LogReg.predict_proba(x_test)
array([[0.82321308, 0.17678692],
[0.40373951, 0.59626049],
[0.60558477, 0.39441523],
[0.94245966, 0.05754034],
[0.84985893, 0.15014107],
[0.18728435, 0.81271565],
[0.25028186, 0.74971814],
[0.3978055 , 0.6021945 ],
[0.70804697, 0.29195303],
[0.92056678, 0.07943322],
[0.3743838 , 0.6256162 ],
[0.67864754, 0.32135246],
[0.75821473, 0.24178527],
[0.88391211, 0.11608789],
[0.77154317, 0.22845683],
[0.55752261, 0.44247739],
[0.92892806, 0.07107194],
[0.93518317, 0.06481683],
[0.65653584, 0.34346416],
[0.82321308, 0.17678692],
[0.91874104, 0.08125896],
[0.118094 , 0.881906 ],
[0.86495191, 0.13504809],
[0.10138955, 0.89861045],
[0.81958824, 0.18041176],
[0.96736015, 0.03263985],
[0.9130323 , 0.0869677 ],
[0.36505302, 0.63494698],
[0.9241062 , 0.0758938 ],
[0.84625216, 0.15374784],
[0.48283594, 0.51716406],
[0.60480059, 0.39519941],
[0.19465748, 0.80534252],
[0.80032396, 0.19967604],
[0.60480059, 0.39519941],
[0.17453915, 0.82546085],
[0.50926883, 0.49073117],
[0.21545788, 0.78454212],
[0.23219685, 0.76780315],
[0.47839638, 0.52160362],
[0.05456151, 0.94543849],
[0.4156908 , 0.5843092 ],
[0.08865759, 0.91134241],
[0.34191125, 0.65808875],
[0.23219685, 0.76780315],
[0.24007614, 0.75992386],
[0.05350511, 0.94649489],
[0.93972035, 0.06027965],
[0.6628199 , 0.3371801 ],
[0.90051653, 0.09948347],
[0.25494693, 0.74505307],
[0.13462671, 0.86537329],
[0.26894863, 0.73105137],
[0.96493672, 0.03506328],
[0.93972035, 0.06027965],
[0.44787585, 0.55212415],
[0.92235496, 0.07764504],
[0.95339818, 0.04660182],
[0.75866386, 0.24133614],
[0.75821473, 0.24178527],
[0.36466091, 0.63533909],
[0.9241062 , 0.0758938 ],
[0.92056678, 0.07943322],
[0.81128902, 0.18871098],
[0.74010796, 0.25989204],
[0.37600706, 0.62399294],
[0.20319182, 0.79680818],
[0.04852806, 0.95147194],
[0.13462671, 0.86537329],
[0.94508177, 0.05491823],
[0.08642183, 0.91357817],
[0.92056678, 0.07943322],
[0.10366311, 0.89633689],
[0.89600078, 0.10399922],
[0.91309258, 0.08690742],
[0.9657632 , 0.0342368 ],
[0.9130323 , 0.0869677 ],
[0.88391211, 0.11608789],
[0.12559105, 0.87440895],
[0.10335856, 0.89664144],
[0.21895463, 0.78104537],
[0.88650007, 0.11349993],
[0.30989833, 0.69010167],
[0.9241062 , 0.0758938 ],
[0.68401238, 0.31598762],
[0.67918193, 0.32081807],
[0.88391211, 0.11608789],
[0.91111165, 0.08888835],
[0.91322676, 0.08677324],
[0.84975293, 0.15024707],
[0.92727933, 0.07272067],
[0.41137752, 0.58862248],
[0.12352188, 0.87647812],
[0.32543235, 0.67456765],
[0.31520742, 0.68479258],
[0.13753152, 0.86246848],
[0.89130479, 0.10869521],
[0.07040303, 0.92959697],
[0.94759103, 0.05240897],
[0.46012544, 0.53987456],
[0.95284617, 0.04715383],
[0.5075382 , 0.4924618 ],
[0.89130479, 0.10869521],
[0.90292346, 0.09707654],
[0.89130479, 0.10869521],
[0.61147104, 0.38852896],
[0.91309258, 0.08690742],
[0.93830535, 0.06169465],
[0.28916704, 0.71083296],
[0.81128902, 0.18871098],
[0.14968075, 0.85031925],
[0.90485703, 0.09514297],
[0.88888767, 0.11111233],
[0.69063218, 0.30936782],
[0.8763524 , 0.1236476 ],
[0.71564929, 0.28435071],
[0.91111165, 0.08888835],
[0.21924413, 0.78075587],
[0.84015292, 0.15984708],
[0.79299019, 0.20700981],
[0.6728628 , 0.3271372 ],
[0.91497443, 0.08502557],
[0.82321308, 0.17678692],
[0.96234036, 0.03765964],
[0.92056678, 0.07943322],
[0.65727554, 0.34272446],
[0.45380545, 0.54619455],
[0.90485703, 0.09514297],
[0.9241062 , 0.0758938 ],
[0.9241062 , 0.0758938 ],
[0.25494693, 0.74505307],
[0.96409103, 0.03590897],
[0.76759616, 0.23240384],
[0.86495191, 0.13504809],
[0.67323514, 0.32676486],
[0.06420273, 0.93579727],
[0.47223412, 0.52776588],
[0.86495191, 0.13504809],
[0.94759103, 0.05240897],
[0.92582112, 0.07417888],
[0.90270837, 0.09729163],
[0.3935528 , 0.6064472 ],
[0.77154317, 0.22845683],
[0.86495191, 0.13504809],
[0.43550251, 0.56449749],
[0.09886755, 0.90113245],
[0.92582112, 0.07417888],
[0.14309623, 0.85690377],
[0.30989833, 0.69010167],
[0.56887211, 0.43112789],
[0.94880468, 0.05119532],
[0.86233075, 0.13766925],
[0.85907445, 0.14092555],
[0.91105013, 0.08894987],
[0.76271567, 0.23728433],
[0.9130323 , 0.0869677 ],
[0.90696312, 0.09303688],
[0.88416341, 0.11583659],
[0.10834576, 0.89165424],
[0.88642364, 0.11357636],
[0.93830535, 0.06169465],
[0.81253892, 0.18746108],
[0.88391211, 0.11608789],
[0.47839638, 0.52160362],
[0.9657632 , 0.0342368 ],
[0.36861451, 0.63138549],
[0.83971168, 0.16028832],
[0.74950028, 0.25049972],
[0.47242338, 0.52757662],
[0.20279521, 0.79720479],
[0.46626928, 0.53373072],
[0.81253892, 0.18746108],
[0.30411983, 0.69588017],
[0.72065067, 0.27934933],
[0.39413803, 0.60586197],
[0.51544237, 0.48455763],
[0.91879772, 0.08120228],
[0.84667869, 0.15332131],
[0.6628199 , 0.3371801 ],
[0.28916704, 0.71083296],
[0.89600078, 0.10399922],
[0.25028186, 0.74971814],
[0.39946546, 0.60053454],
[0.68915076, 0.31084924],
[0.88391211, 0.11608789],
[0.23219685, 0.76780315]])
plt.figure(figsize=(5,5))
sns.barplot(x="Sex_female",y="Survived",data=df_train_dummied)
<matplotlib.axes._subplots.AxesSubplot at 0x7f9b55ee9ef0>
Above shows that more number of Females survived and bottomone shows more male did not survive.
sns.barplot(x="Sex_male",y="Survived",data=df_train_dummied)
<matplotlib.axes._subplots.AxesSubplot at 0x7f9b55e4eef0>
On a cautionary Note, bottom I tried to plot Class vs Suriveval plot, but it is Lurking variable syndrome. Not to be accounted.
sns.barplot(x="Class",y="Survived",data=df_train_dummied)
<matplotlib.axes._subplots.AxesSubplot at 0x7f9b55957e80>
sns.barplot(x="Class",y="Sex_male",data=df_train_dummied)
<matplotlib.axes._subplots.AxesSubplot at 0x7f9b55445b00>
Even Cost vs Survival,
sns.barplot(x="Class",y="Cost",data=df_train_dummied)
<matplotlib.axes._subplots.AxesSubplot at 0x7f9b546a7be0>
