University education is becoming a critical pillar of social and economic life in the twenty-first century. It is crucial not only in the educational process but also in assuring two vital things: a great job and financial stability. Predicting university entrance, on the other hand, might be extremely challenging because pupils are unaware of the admission standards.
So in this tutorial, we will be building our own University Admission Prediction model using Python Programming Language.
Introduction to Dataset
When applying for Master’s degrees in foreign countries, there are several variables to consider. You must have a decent GRE score, a sop (statement of purpose), or a letter of reference, among other things. If you are not from an English-speaking nation, you need also to submit a TOEFL score.
You can access the dataset here. The dataset includes the following attributes:
- GRE Scores ( out of 340 )
- TOEFL Scores ( out of 120 )
- University Rating ( out of 5 )
- Statement of Purpose and Letter of Recommendation Strength ( out of 5 )
- Undergraduate GPA ( out of 10 )
- Research Experience ( either 0 or 1 )
- Chance of Admit ( ranging from 0 to 1 )
Implementing University Admission Protection in Python
We would be dividing the whole code implementation into a number of steps as mentioned below:
Step 1: Importing Necessary Modules/Libraries
import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
from keras.models import Sequential
from keras.layers import Dense ,Dropout,BatchNormalization
from keras.layers import Dense
from keras.wrappers.scikit_learn import KerasRegressor
Step 2: Loading Dataset Into the Program
df = pd.read_csv('Admission_Predict.csv')
df.head()
Step 3: Data Pre-processing and Data Splitting
Before building our main model, we would require some pre-processing which involves dropping any column which is not necessary for the model.
Here the column ‘Serial No.’ isn’t necessary for admission prediction so we drop it out of the data.
df=df.drop("Serial No.",axis=1)
After this we would be dividing the dataset into X and Y sub-datasets where X will have all the information and Y will include the final probability.
Y=np.array(df[df.columns[-1]])
X=np.array(df.drop(df.columns[-1],axis=1))
Now, the next step is to split the dataset into training and testing datasets using the 80:20 train test split rule where 80% of the data is used for training and the rest 20% is used for testing.
X_train, X_test, y_train, y_test = train_test_split(X,Y, test_size=0.2, random_state=0)
The pre-processing will also involve normalizing the training dataset which can be achieved through the code mentioned below.
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
X_train=scaler.fit_transform(X_train)
X_test=scaler.fit_transform(X_test)
Step 3: Building the Model
The code mentioned below is the main function that describes the whole model involving the declaration of the model and adding layers to the model.
The function also involves compilation of the model and computation of the loss.
def baseline_model():
model = Sequential()
model.add(Dense(16, input_dim=7, activation='relu'))
model.add(Dense(16, input_dim=7, activation='relu'))
model.add(Dense(16, input_dim=7, activation='relu'))
model.add(Dense(16, input_dim=7, activation='relu'))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adam')
return model
Step 4: Training of the Model
The next step is to create out the model object and train the same on the training dataset as mentioned in the code below. You can keep the number of epochs according to your own preference.
estimator = KerasRegressor(build_fn=baseline_model, epochs=50, batch_size=3, verbose=1)
estimator.fit(X_train,y_train)
The output of the training is as follows:
Epoch 1/50
107/107 [==============================] - 1s 3ms/step - loss: 0.1087
Epoch 2/50
107/107 [==============================] - 0s 4ms/step - loss: 0.0065
Epoch 3/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0057
Epoch 4/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0052
Epoch 5/50
107/107 [==============================] - 0s 4ms/step - loss: 0.0049
Epoch 6/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0050
Epoch 7/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0047
Epoch 8/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0049
Epoch 9/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0044
Epoch 10/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0043
Epoch 11/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0044
Epoch 12/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0044
Epoch 13/50
107/107 [==============================] - 0s 4ms/step - loss: 0.0043
Epoch 14/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0041
Epoch 15/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0043
Epoch 16/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0042
Epoch 17/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0040
Epoch 18/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0043
Epoch 19/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0039
Epoch 20/50
107/107 [==============================] - 0s 4ms/step - loss: 0.0040
Epoch 21/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0039
Epoch 22/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0042
Epoch 23/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0040
Epoch 24/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0038
Epoch 25/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0042
Epoch 26/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0038
Epoch 27/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0040
Epoch 28/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0042
Epoch 29/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0039
Epoch 30/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0037
Epoch 31/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0038
Epoch 32/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0043
Epoch 33/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0040
Epoch 34/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0037
Epoch 35/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0039
Epoch 36/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0037
Epoch 37/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0038
Epoch 38/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0036
Epoch 39/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0036
Epoch 40/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0036
Epoch 41/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0037
Epoch 42/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0037
Epoch 43/50
107/107 [==============================] - 0s 4ms/step - loss: 0.0036
Epoch 44/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0037
Epoch 45/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0037
Epoch 46/50
107/107 [==============================] - 0s 4ms/step - loss: 0.0038
Epoch 47/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0036
Epoch 48/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0037
Epoch 49/50
107/107 [==============================] - 0s 4ms/step - loss: 0.0037
Epoch 50/50
107/107 [==============================] - 0s 3ms/step - loss: 0.0034
<keras.callbacks.History at 0x7f10c0173e10>
[19]
0s
Step 5: Testing the Model
Now, let us try to predict the values for the testing dataset and match them with the original values.
prediction = estimator.predict(X_test)
print("ORIGINAL DATA")
print(y_test)
print()
print("PREDICTED DATA")
print(prediction)
The output looks somewhat like this:
ORIGINAL DATA
[0.71 0.7 0.79 0.73 0.72 0.48 0.77 0.71 0.9 0.94 0.58 0.89 0.72 0.57
0.78 0.42 0.64 0.84 0.63 0.72 0.9 0.83 0.57 0.47 0.85 0.67 0.44 0.54
0.92 0.62 0.68 0.73 0.73 0.61 0.55 0.74 0.64 0.89 0.73 0.95 0.71 0.72
0.75 0.76 0.86 0.7 0.39 0.79 0.61 0.64 0.71 0.8 0.61 0.89 0.68 0.79
0.78 0.52 0.76 0.88 0.74 0.49 0.65 0.59 0.87 0.89 0.81 0.9 0.8 0.76
0.68 0.87 0.68 0.64 0.91 0.61 0.69 0.62 0.93 0.43]
PREDICTED DATA
[0.64663166 0.6811929 0.77187485 0.59903866 0.70518774 0.5707331
0.6844891 0.6232987 0.8559068 0.9225058 0.50917023 0.9055291
0.6913604 0.40199894 0.8595592 0.6155516 0.5891675 0.793468
0.5415057 0.7054745 0.8786436 0.8063141 0.55548865 0.3587063
0.77944946 0.5391258 0.43374807 0.62050253 0.90883577 0.6109837
0.64160395 0.7341113 0.73316455 0.5032365 0.7664028 0.76009744
0.59858805 0.86267006 0.60282356 0.94984144 0.7196544 0.63529354
0.7032968 0.8164513 0.8044792 0.6359613 0.54865533 0.6914524
0.589018 0.55952907 0.6446153 0.77345765 0.6449453 0.8998446
0.68746895 0.74362046 0.71107167 0.73258513 0.7594558 0.8374823
0.7504637 0.4027493 0.61975926 0.46762955 0.8579673 0.814696
0.7111042 0.8707262 0.7539967 0.7515583 0.5506843 0.8436626
0.8139006 0.5593421 0.933276 0.61958474 0.6084135 0.63294107
0.9234169 0.44476634]
You can see that the values do match to some extent. But let’s make sure, we compute the mean error as well.
Step 6: Computing Mean Error
from sklearn.metrics import accuracy_score
train_error = np.abs(y_test - prediction)
mean_error = np.mean(train_error)
print("Mean Error: ",mean_error)
The mean error comes out to be 0.0577927375137806 which is good enough to say that our results are pretty accurate.
Conclusion
Congratulations! You just learned about making your own University Admission Predictor. Hope you enjoyed it! 😇
Liked the tutorial? In any case, I would recommend you to have a look at the tutorials mentioned below:
- Crypto Price Prediction with Python
- Box Office Revenue Prediction in Python – An Easy Implementation
- Stock Price Prediction using Python
- Wine Classification using Python – Easily Explained
Thank you for taking your time out! Hope you learned something new!! 😄