Embed-ML-model-to-Web

This is a simple web based ML mode

Aim of the model is to get the weightage of the elements as input for the type glass and predict the type of glass usage

The Model folder contains the ML program in python which is converted to .sav using joblib python module.

That .sav file is used in django to integrate with the website

The djangoMl folder contains the django files which using put it in your django virtual environment and add the path (urls.py)to this folder

The ml folder contains the template of the website in html format (home.html for input) and (result.html for output)

Adjust the setting.py file in django so that you can integrate the template with the django view.py

Similar way you can add your model and adjust the webiste too.

Important python module

joblib

pip3 install joblib

This module will convert the python Ml model to sav file which can be further used in django-website to predict result

Machine Learning model used

random forest (model.py)

import numpy as np
import pandas as pd
import joblib

dataset = pd.read_csv('glass.csv')

x = dataset.iloc[:,:-1]
y = dataset.iloc[:,9]

from sklearn.model_selection import train_test_split
x_train,x_test,y_train,y_test = train_test_split(x,y,test_size = 0.20,random_state = 42)

from sklearn.preprocessing import StandardScaler
sc_x = StandardScaler()
x_train = sc_x.fit_transform(x_train)
x_test = sc_x.transform(x_test)

from sklearn.ensemble import RandomForestClassifier

clss = RandomForestClassifier(criterion='entropy',n_estimators=300,random_state=42)
clss.fit(x_train,y_train)

print('accuracy is ', clss.score(x_test,y_test)*100,'%')

#saving the model
filename = 'finalized_model.sav'
joblib.dump(clss,filename)

Code Snippet used in django views

view.py

from django.shortcuts import render
import joblib
# Create your views here.
def home_view(request):
    return render(request,"ml/home.html")

def result_view(request):
    clss = joblib.load('model/finalized_model.sav')
    lis = []
    lis.append(request.GET['RI'])
    lis.append(request.GET['Na'])
    lis.append(request.GET['Mg'])
    lis.append(request.GET['Al'])
    lis.append(request.GET['Si'])
    lis.append(request.GET['K'])
    lis.append(request.GET['Ca'])
    lis.append(request.GET['Ba'])
    lis.append(request.GET['Fe'])
    
    ans = clss.predict([lis])
    
    usage = [
        'building windows float processed',
        'building windows non float processed',
        'vehicle windows  float processed',
        'vehicle windows float processed (none in this dataset)',
        'containers',
        'tableware',
        'headlamps'
    ]
    
    return render(request,"ml/result.html",{'ans':usage[ans[0] - 1]})

geek-ninja / embed-ml-model-to-web Goto Github PK