From d0b67c4dd0d9b738b05547c42561be7f73264366 Mon Sep 17 00:00:00 2001
From: mehtataral <75374656+mehtataral@users.noreply.github.com>
Date: Wed, 24 May 2023 19:23:53 +0530
Subject: [PATCH] Add files via upload

---
 5thRegressionMultiReg (1).py          |  70 ++++++++++++++++++
 5thRegressionMultiReg.py              |  77 ++++++++++++++++++++
 Mumbai_regression_housePredicvtion.py |  83 +++++++++++++++++++++
 house_pred_duplicate.csv              |  11 +++
 labelencoder_linearregression.py      |  99 +++++++++++++++++++++++++
 lassoReg.py                           |  67 +++++++++++++++++
 linearRegresssion.py                  |  65 +++++++++++++++++
 linear_reg_train.py                   |  99 +++++++++++++++++++++++++
 linear_regression_final.py            | 100 ++++++++++++++++++++++++++
 multi linear regression.py            |  63 ++++++++++++++++
 multi_house_para.csv                  |   6 ++
 11 files changed, 740 insertions(+)
 create mode 100644 5thRegressionMultiReg (1).py
 create mode 100644 5thRegressionMultiReg.py
 create mode 100644 Mumbai_regression_housePredicvtion.py
 create mode 100644 house_pred_duplicate.csv
 create mode 100644 labelencoder_linearregression.py
 create mode 100644 lassoReg.py
 create mode 100644 linearRegresssion.py
 create mode 100644 linear_reg_train.py
 create mode 100644 linear_regression_final.py
 create mode 100644 multi linear regression.py
 create mode 100644 multi_house_para.csv

diff --git a/5thRegressionMultiReg (1).py b/5thRegressionMultiReg (1).py
new file mode 100644
index 0000000..544a833
--- /dev/null
+++ b/5thRegressionMultiReg (1).py	
@@ -0,0 +1,70 @@
+import pandas as pd
+df=pd.read_csv(open("D:/desktop/AI Training Content/multireg.csv","rb"))
+
+x=df.iloc[:,:-1]
+y=df.iloc[:,-1]
+
+
+import numpy as np
+x=np.array(x)
+y=np.array(y)
+
+
+from sklearn.impute import SimpleImputer
+im=SimpleImputer()
+x=im.fit_transform(x)   #rule: min 2D array
+
+
+from sklearn.preprocessing import MinMaxScaler
+sc=MinMaxScaler()
+x=sc.fit_transform(x)
+
+from sklearn.model_selection import train_test_split
+xtrain,xtest,ytrain,ytest=train_test_split(x,y, test_size=0.2)
+
+
+#now select the algo: as target is of continous type use regression algo: Linear Regression, RandomForest
+# it works on equation of a line: y=bx+a     
+
+#Types: Simple linear regression {1 input column: 1 output column} 2) Multi linear regression{more than 1 input and 1 output}.
+  
+from sklearn.linear_model import LinearRegression
+lr=LinearRegression()
+lr.fit(xtrain,ytrain)
+pred=lr.predict(xtest)
+
+#for regression the evaluation method is completely different.
+#1) visualization
+
+xaxis=np.linspace(1,len(pred),len(pred))
+import matplotlib.pyplot as plt
+plt.plot(xaxis,pred,color='red')
+plt.plot(xaxis,ytest,color='blue')
+plt.show()
+
+#2) how much good??? to define it in a number RMSE calculate
+from sklearn.metrics import mean_squared_error
+res=np.sqrt(mean_squared_error(pred,ytest))
+
+
+
+from sklearn.ensemble import RandomForestRegressor
+rf=RandomForestRegressor()
+rf.fit(xtrain,ytrain)
+predrf=rf.predict(xtest)
+
+#for regression the evaluation method is completely different.
+#1) visualization
+
+xaxis=np.linspace(1,len(predrf),len(predrf))
+import matplotlib.pyplot as plt
+plt.plot(xaxis,predrf,color='red')
+plt.plot(xaxis,ytest,color='blue')
+plt.show()
+
+#2) how much good??? to define it in a number RMSE calculate
+from sklearn.metrics import mean_squared_error
+resrf=np.sqrt(mean_squared_error(predrf,ytest))
+
+
+
diff --git a/5thRegressionMultiReg.py b/5thRegressionMultiReg.py
new file mode 100644
index 0000000..9219379
--- /dev/null
+++ b/5thRegressionMultiReg.py
@@ -0,0 +1,77 @@
+import pandas as pd
+# df=pd.read_csv(open("D:/desktop/AI Training Content/multireg.csv","rb"))
+df=pd.read_csv(open("C:/Users/Admin/Downloads/Churn_Modelling.csv","rb"))
+
+
+# x=df.iloc[:,:-1]
+# y=df.iloc[:,-1]
+x=df.iloc[:,3:-1]
+y=df.iloc[:,-1]
+
+import numpy as np
+x=np.array(x)
+y=np.array(y)
+
+from sklearn.preprocessing import LabelEncoder
+lbl=LabelEncoder()
+x[:,1]=lbl.fit_transform(x[:,1])
+x[:,2]=lbl.fit_transform(x[:,2])
+
+from sklearn.impute import SimpleImputer
+im=SimpleImputer()
+x=im.fit_transform(x)   #rule: min 2D array
+
+
+from sklearn.preprocessing import MinMaxScaler
+sc=MinMaxScaler()
+x=sc.fit_transform(x)
+
+from sklearn.model_selection import train_test_split
+xtrain,xtest,ytrain,ytest=train_test_split(x,y, test_size=0.2)
+
+
+#now select the algo: as target is of continous type use regression algo: Linear Regression, RandomForest
+# it works on equation of a line: y=bx+a     
+
+#Types: Simple linear regression {1 input column: 1 output column} 2) Multi linear regression{more than 1 input and 1 output}.
+  
+from sklearn.linear_model import LinearRegression
+lr=LinearRegression()
+lr.fit(xtrain,ytrain)
+pred=lr.predict(xtest)
+
+#for regression the evaluation method is completely different.
+#1) visualization
+
+xaxis=np.linspace(1,len(pred),len(pred))
+import matplotlib.pyplot as plt
+plt.plot(xaxis,pred,color='red')
+plt.plot(xaxis,ytest,color='blue')
+plt.show()
+
+#2) how much good??? to define it in a number RMSE calculate
+from sklearn.metrics import mean_squared_error
+res=np.sqrt(mean_squared_error(pred,ytest))
+
+
+
+from sklearn.ensemble import RandomForestRegressor
+rf=RandomForestRegressor()
+rf.fit(xtrain,ytrain)
+predrf=rf.predict(xtest)
+
+#for regression the evaluation method is completely different.
+#1) visualization
+
+xaxis=np.linspace(1,len(predrf),len(predrf))
+import matplotlib.pyplot as plt
+plt.plot(xaxis,predrf,color='red')
+plt.plot(xaxis,ytest,color='blue')
+plt.show()
+
+#2) how much good??? to define it in a number RMSE calculate
+from sklearn.metrics import mean_squared_error
+resrf=np.sqrt(mean_squared_error(predrf,ytest))
+
+
+
diff --git a/Mumbai_regression_housePredicvtion.py b/Mumbai_regression_housePredicvtion.py
new file mode 100644
index 0000000..d76941d
--- /dev/null
+++ b/Mumbai_regression_housePredicvtion.py
@@ -0,0 +1,83 @@
+import numpy as np
+import pandas as pd
+
+
+df =pd.read_csv(r"C:\Users\user\Documents\Taral\DATA_SET\DATA_SET\Regression Data Set\Mumbai1.csv")
+print(df)
+df.info()
+df.drop("Unnamed: 0",axis ="columns",inplace =True)
+
+x = df.iloc[:,1:]#input
+print(x)
+x.info()
+x.corr()
+y =df.iloc[:,0]#output
+print(y)
+
+corr = x.corr()
+#training and testing data
+
+from sklearn.preprocessing import LabelEncoder
+lb = LabelEncoder()
+x.Location = lb.fit_transform(x.Location)
+print(x)
+
+from sklearn.model_selection import train_test_split
+x_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.3,random_state=500)
+
+
+print(len(x_train))
+print(len(x_test))
+print(len(y_train))
+print(len(y_test))
+
+
+print(x_train)
+print(x_test)
+print(y_train)
+print(y_test)
+
+
+from sklearn.linear_model import LinearRegression
+lbl =LinearRegression()
+lbl.fit(x_train,y_train)
+lbl.predict(x_test)
+lbl.score(x_test,y_test)
+
+from sklearn.linear_model import Lasso,Ridge
+las1 = Lasso()
+
+las1.fit(x_train,y_train)
+
+las1.predict(x_test)
+las1.score(x_test,y_test)
+
+
+las1 = Lasso(alpha =1000,selection="random")
+
+las1.fit(x_train,y_train)
+
+las1.predict(x_test)
+las1.score(x_test,y_test)
+   
+
+rid1 = Ridge()
+
+rid1.fit(x_train,y_train)
+
+predict1=rid1.predict(x_test)
+rid1.score(x_test,y_test)
+
+# dff = pd.DataFrame(predict1)
+# dff
+# dff["actual_value"] =y_test
+# dff
+
+rid2 = Ridge(alpha =50)
+
+rid2.fit(x_train,y_train)
+
+rid2.predict(x_test)
+rid2.score(x_test,y_test)
+rid2.score(x_train,y_train)
+
diff --git a/house_pred_duplicate.csv b/house_pred_duplicate.csv
new file mode 100644
index 0000000..508cda9
--- /dev/null
+++ b/house_pred_duplicate.csv
@@ -0,0 +1,11 @@
+area,price
+3200,61000000
+2600,55000000
+2600,55000000
+3000,56500000
+3200,61000000
+3600,59500000
+4000,76000000
+3000,
+4011,
+2600,55000000
diff --git a/labelencoder_linearregression.py b/labelencoder_linearregression.py
new file mode 100644
index 0000000..5617c9f
--- /dev/null
+++ b/labelencoder_linearregression.py
@@ -0,0 +1,99 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Fri Feb 25 12:06:40 2022
+Label Encoder Multilinear Regression
+
+@author: user
+"""
+import numpy as np
+import pandas as pd
+import matplotlib.pyplot as plt
+
+df =pd.read_csv(r"C:\Users\user\Documents\Taral\DATA_SET\DATA_SET\Regression Data Set\Basic Regression Data\multi_house_dummy.csv")
+print(df)
+df = df.reindex(columns=["area","town","price"])
+print(df)
+
+x = df.iloc[:,:-1]
+print(x)
+
+y =df.iloc[:,-1:]
+print(y)
+
+plt.scatter(x.area, y)
+plt.show()
+
+#LabelEncoder converts strings inot numbers
+from sklearn.preprocessing import LabelEncoder
+lbl =LabelEncoder()
+x.town = lbl.fit_transform(x.town)
+print(x)
+
+from sklearn.linear_model import LinearRegression
+reg = LinearRegression()
+reg.fit(x,y)
+reg.predict(x)
+
+from sklearn.preprocessing import PolynomialFeatures
+from sklearn.linear_model import LinearRegression
+
+pr = PolynomialFeatures(degree =3)
+x_poly = pr.fit_transform(x)
+reg = LinearRegression()
+reg.fit(x_poly,y)
+
+#lasso regression 
+#ridge regression
+#polynomail regression --> scatter
+
+print(reg.coef_)
+print(reg.intercept_)
+
+reg.predict(x_poly)
+reg.score(x_poly,y)
+
+reg.predict([[3600]])
+reg.predict([[2600]])
+reg.predict([[4600]])
+
+reg.predict([[1,2600]])
+reg.predict([[2,2600]])
+reg.score(x,y)
+
+plt.scatter(x.area, y)
+plt.plot(x.area,reg.predict(x),'rD:')
+plt.show()
+
+
+from sklearn.linear_model import Ridge,Lasso
+from sklearn.model_selection import GridSearchCV
+model1=Ridge(alpha=5)
+model1.fit(x,y)
+model1.score(x,y)
+print(model1.coef_)#coefiicent
+print(model1.intercept_)
+model1.predict([[2600]])
+model1.predict([[3600]])
+model1.predict([[4600]])
+model1.predict([[2000]])
+
+model2=Lasso(alpha=6)
+model2.fit(x,y)
+model2.score(x,y)
+print(model2.coef_)
+print(model2.intercept_)
+
+
+m1=Ridge()
+parameters={'alpha':[1e-15,1e-18,1e-8,1e-4,1e-3,1e-2,1,5,10,20]}
+
+lasso_regressor=GridSearchCV(m1,parameters,cv=5)
+lasso_regressor.fit(x,y)
+lasso_regressor.score(x,y)
+
+print(model2.coef_)
+
+# from sklearn.metrics import mean_squared_error
+# mse = mean_squared_error(y,ypred)
+# rmse= np.sqrt(mse)
+# 
\ No newline at end of file
diff --git a/lassoReg.py b/lassoReg.py
new file mode 100644
index 0000000..a8ba861
--- /dev/null
+++ b/lassoReg.py
@@ -0,0 +1,67 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Fri Feb 25 12:26:46 2022
+
+@author: user
+"""
+
+# -*- coding: utf-8 -*-
+"""
+Created on Mon Feb 10 12:29:51 2020
+LASSO
+@author: Trainer 1
+"""
+x =(1,2,3)
+lst =list(x)
+print(lst)
+lst.append(55)
+lst
+x =tuple(lst)
+import pandas as pd
+import numpy as np
+df=pd.read_csv('Advertising.csv')
+df.head()
+df.drop('SR',axis=1,inplace=True)
+
+from sklearn.linear_model import LinearRegression
+x=df.drop('SALE',axis=1)
+y=pd.DataFrame(df.SALE)
+model1=LinearRegression()
+model1.fit(x,y)
+print(model1.coef_)
+ki7i
+
+from sklearn.linear_model import Ridge,Lasso
+from sklearn.model_selection import GridSearchCV
+model2=Ridge(alpha=700)
+model2.fit(x,y)
+print(model2.coef_)
+
+
+model3=Lasso(alpha=6)
+model3.fit(x,y)
+model3.coef_
+m1=Ridge()
+parameters={'alpha':[1e-15,1e-18,1e-8,1e-4,1e-3,1e-2,1,5,10,20]}
+
+lasso_regressor=GridSearchCV(m1,parameters,cv=5)
+lasso_regressor.fit(x,y)
+print(model3.coef_)
+print(lasso_regressor.best_params_)
+print(lasso_regressor.best_score_)
+
+
+89,102,88,0  :Lasso n rigid 
+y=89*555676+0*23454+
+
+higer no feature n data : processing time n compution cost
+
+when we say Processing Time  incerse : n no calct happen backen  
+foreg :
+    y=m1x1+m2x2+m3x3+m4x4 +c
+    m1=9,m2=6,m3=4,m4=2, x1=100,x2=350,x3=55,x4=10
+what if i have  m1=220,m2=300,m4=267,m3=200 then processing time then we need to find a method to decrese the values  how ???
+ Lasso n Rigid : decrese the values 
+ 
+
+Featureselection : chisquare test 
diff --git a/linearRegresssion.py b/linearRegresssion.py
new file mode 100644
index 0000000..18fad53
--- /dev/null
+++ b/linearRegresssion.py
@@ -0,0 +1,65 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Thu Feb 24 17:10:59 2022
+
+@author: user
+"""
+import numpy as np
+import pandas as pd
+import matplotlib.pyplot as plt
+
+x ={
+    "area":[2600,3000,3200,3600,4000],
+    "price":["55000000","56500000","61000000","59500000","76000000"]
+    }
+
+print(x)
+df =pd.DataFrame(x)
+print(df)
+# y ={
+#     "area":[2600,3000,3200,3600,4000,2800,3400,3800,3900,4900]
+#     }
+
+# print(y)
+# df =pd.DataFrame(y)
+# print(df)
+# df.to_csv(r"C:\Users\user\Desktop\DS\DATA_SET\house_pred.csv",index =False)
+# df.to_csv(r"C:\Users\user\Desktop\DS\DATA_SET\house_area.csv",index =False)
+
+
+# df =pd.read_csv(r"C:\Users\user\Desktop\DS\DATA_SET\house_pred.csv")
+# print(df)
+
+x = df.iloc[:,0]
+y = df.iloc[:,-1]
+plt.xlabel("Price")
+plt.ylabel("Area")
+plt.scatter(df.price,df.area,marker="+")
+plt.plot(df.price,df.area)
+plt.show()
+
+from sklearn.linear_model import LinearRegression
+reg =LinearRegression()
+# reg.fit([[x],[y]])
+reg.fit(df[['area']],df.price)
+# reg.fit(x[x['area']],y.price) #not valid ..used xtrain and ytrain
+reg.coef_
+reg.intercept_
+x =reg.predict([[3300]])
+reg.predict([[2600]])
+reg.predict([[4500]])
+
+
+x =reg.predict(df)#only area is mentioned in df 
+df['price'] = x #store that price values in price
+print(df)    
+
+plt.xlabel("Price")
+plt.ylabel("Area")
+plt.scatter(df.price,df.area,marker="+")
+plt.plot(df.area,reg.predict(df[['area']]),color ="green",marker = "o")
+plt.show()
+
+# from sklearn.metrics import accuracy_score
+# acc =accuracy_score((2600),(4500))#not valid ..used xtrain and ytrain
+# print(acc)
diff --git a/linear_reg_train.py b/linear_reg_train.py
new file mode 100644
index 0000000..b2994b2
--- /dev/null
+++ b/linear_reg_train.py
@@ -0,0 +1,99 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Fri Feb 25 12:38:01 2022
+train and test Multilinear Regression
+
+@author: user
+"""
+
+import numpy as np
+import pandas as pd
+
+
+df =pd.read_csv(r"C:\Users\user\Documents\Taral\DATA_SET\DATA_SET\Regression Data Set\Basic Regression Data\multi_house_dummy.csv")
+print(df)
+
+x = df.iloc[:,:-1]
+print(x)
+
+y =df.iloc[:,-1]
+print(y)
+
+a = pd.get_dummies(x.town)
+print(a)
+
+new_df = pd.concat([x,a],axis="columns")
+print(new_df)
+
+new_df =new_df.drop("town",axis ="columns")
+print(new_df)
+
+#training and testing data
+
+from sklearn.model_selection import train_test_split 
+x_train,x_test,y_train,y_test = train_test_split(new_df,y,test_size=0.3,random_state=5)
+
+
+print(len(x_train))
+print(len(x_test))
+print(len(y_train))
+print(len(y_test))
+
+
+print(x_train)
+print(x_test)
+print(y_train)
+print(y_test)
+
+
+from sklearn.linear_model import LinearRegression
+lbl =LinearRegression()
+lbl.fit(x_train,y_train)
+
+
+lbl.predict(x_test)
+y_test
+lbl.score(x_test,y_test)
+
+lbl.score(x_train,y_train)
+
+
+
+
+
+from sklearn.linear_model import Lasso,Ridge
+las1 = Lasso()
+
+las1.fit(x_train,y_train)
+
+las1.predict(x_test)
+las1.score(x_test,y_test)
+
+
+
+las1 = Lasso(alpha =100)
+
+las1.fit(x_train,y_train)
+
+las1.predict(x_test)
+las1.score(x_test,y_test)
+ 
+
+rid1 = Ridge()
+
+rid1.fit(x_train,y_train)
+
+rid1.predict(x_test)
+rid1.score(x_test,y_test)
+
+
+
+rid2 = Ridge(alpha =10)
+
+rid2.fit(x_train,y_train)
+
+rid2.predict(x_test)
+rid2.score(x_test,y_test)
+
+
+
diff --git a/linear_regression_final.py b/linear_regression_final.py
new file mode 100644
index 0000000..ec633d0
--- /dev/null
+++ b/linear_regression_final.py
@@ -0,0 +1,100 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Tue Mar 22 19:52:36 2022
+
+@author: user
+"""
+
+import numpy as np
+import pandas as pd
+
+
+df =pd.read_csv(r"C:\Users\user\Desktop\DS\DATA_SET\Regression Data Set\CAR DETAILS FROM CAR DEKHO.csv")
+print(df)
+df.info()
+
+df = df.reindex(columns = ["name","year","km_driven","fuel","seller_type","transmission","owner","selling_price"])
+
+x = df.iloc[:,:-1]
+print(x)
+x.info()
+x.corr()
+y =df.iloc[:,-1]
+print(y)
+
+a = pd.get_dummies(x.fuel)
+print(a)
+
+b =pd.get_dummies(x.seller_type)
+print(b)
+
+c =pd.get_dummies(x.transmission)
+print(c)
+
+d =pd.get_dummies(x.owner)
+print(d)
+
+
+new_df = pd.concat([x,a,b,c,d],axis="columns")
+print(new_df)
+
+new_df =new_df.drop(["fuel","seller_type","transmission","owner","name"],axis ="columns")
+print(new_df)
+new_df.corr()
+#training and testing data
+
+from sklearn.model_selection import train_test_split
+x_train,x_test,y_train,y_test = train_test_split(new_df,y,test_size=0.3,random_state=500)
+
+
+print(len(x_train))
+print(len(x_test))
+print(len(y_train))
+print(len(y_test))
+
+
+print(x_train)
+print(x_test)
+print(y_train)
+print(y_test)
+
+
+from sklearn.linear_model import LinearRegression
+lbl =LinearRegression()
+lbl.fit(x_train,y_train)
+lbl.predict(x_test)
+lbl.score(x_test,y_test)
+
+from sklearn.linear_model import Lasso,Ridge
+las1 = Lasso()
+
+las1.fit(x_train,y_train)
+
+las1.predict(x_test)
+las1.score(x_test,y_test)
+
+
+
+las1 = Lasso(alpha =100)
+
+las1.fit(x_train,y_train)
+
+las1.predict(x_test)
+las1.score(x_test,y_test)
+ 
+
+rid1 = Ridge()
+
+rid1.fit(x_train,y_train)
+
+rid1.predict(x_test)
+rid1.score(x_test,y_test)
+
+
+
+rid2 = Ridge(alpha =50)
+
+rid2.fit(x_train,y_train)
+
+rid2.predict(x_test)
+rid2.score(x_test,y_test)
diff --git a/multi linear regression.py b/multi linear regression.py
new file mode 100644
index 0000000..c915aa7
--- /dev/null
+++ b/multi linear regression.py	
@@ -0,0 +1,63 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Thu Feb 24 19:04:12 2022
+
+@author: user
+"""
+import pandas as pd
+import numpy as np
+# x ={
+#     "area":[2600,3000,3200,3600,4000],
+#     "bedroom":[3,4,np.nan,3,5],
+#     "age_house":[20,15,np.nan,30,8],
+#     "price":["55000000","56500000","61000000","59500000","76000000"]
+#     }
+# print(x)
+
+# df =pd.DataFrame(x)
+# print(df)
+# df.to_csv(r"C:\Users\user\Desktop\DS\DATA_SET\multi_house_para.csv",index =False)
+
+df =pd.read_csv(r"C:\Users\user\Documents\Taral\DataScience\Regression\multi_house_para.csv")
+print(df)
+
+
+df["bedroom"] =df['bedroom'].fillna(df.bedroom.mean())
+print(df)
+df["age_house"] = df.age_house.fillna(df.age_house.mean())
+print(df)
+
+
+df.bedroom = df.bedroom.astype(int)
+df.age_house = df.age_house.astype(int)
+print(df)
+
+x =df.iloc[:,:-1]
+y =df.iloc[:,-1]
+print(x)
+print(y)
+from sklearn.linear_model import LinearRegression
+reg = LinearRegression()
+reg.fit(x,y)
+reg.predict([[2600,3,20]])#area, bedroom , age_house
+reg.predict([[2600,3,10]])
+reg.predict([[2600,3,1]]) 
+reg.predict([[600,1,30]])
+
+print(reg.score(x,y))
+
+reg.fit(df[["area","bedroom","age_house"]],df.price)
+reg.predict([[2600,3,20]])#area, bedroom , age_house
+reg.predict([[2600,3,10]])
+reg.predict([[2600,3,1]])
+
+reg.predict([[3600,3,30]])
+
+
+
+
+
+
+
+
+
diff --git a/multi_house_para.csv b/multi_house_para.csv
new file mode 100644
index 0000000..8932140
--- /dev/null
+++ b/multi_house_para.csv
@@ -0,0 +1,6 @@
+area,bedroom,age_house,price
+2600,3,20,55000000
+3000,4,15,56500000
+3200,,,61000000
+3600,3,30,59500000
+4000,5,8,76000000