A diabetes detection machine learning project involves using data and algorithms to train a model to accurately predict the likelihood of an individual having diabetes based on various features such as Glucose, age, and blood Pressure.
Diabetes detection machine learning project is a system that uses algorithms, statistical models and historical data to predict the likelihood of an individual having diabetes. The goal is to use features like Glucose, age, and Blood Pressure to identify individuals who have diabetes or are at risk of developing it. Machine learning models are trained using this data and can then be used to make predictions on new individuals.
This dataset is originally from the National Institute of Diabetes and Digestive and Kidney Diseases. The objective is to predict based on diagnostic measurements whether a patient has diabetes.
Several constraints were placed on the selection of these instances from a larger database. In particular, all patients here are females at least 21 years old of Pima Indian heritage.
- Pregnancies: Number of times pregnant
- Glucose: Plasma glucose concentration a 2 hours in an oral glucose tolerance test
- BloodPressure: Diastolic blood pressure (mm Hg)
- SkinThickness: Triceps skin fold thickness (mm)
- Insulin: 2-Hour serum insulin (mu U/ml)
- BMI: Body mass index (weight in kg/(height in m)^2)
- DiabetesPedigreeFunction: Diabetes pedigree function
- Age: Age (years)
- Outcome: Class variable (0 or 1)
(a) Original owners: National Institute of Diabetes and Digestive and Kidney Diseases (b) Donor of database: Vincent Sigillito ([email protected]) Research Center, RMI Group Leader Applied Physics Laboratory The Johns Hopkins University Johns Hopkins Road Laurel, MD 20707 (301) 953-6231 (c) Date received: 9 May 1990
The process of feature selection in machine learning is used to identify and select the most relevant features from a dataset to improve the performance and efficiency of a machine learning model. It is an important step in the model building process as it can help to reduce overfitting, increase model interpretability, and improve the accuracy of predictions.
A correlation matrix can be used to identify highly correlated features, which can then be removed or consolidated. Highly correlated features can cause a problem in machine learning models as they can introduce multicollinearity, which can lead to unstable and unreliable model estimates.
Outlier treatment is the process of identifying and handling extreme values or observations that are significantly different from other observations in a dataset. Outliers can have a significant impact on the results of data analysis and modeling, and can skew the mean and standard deviation of a dataset.
-
Visualization: Outliers can be identified by creating visualizations such as box plots, scatter plots, and histograms to identify observations that fall outside of the typical range.
-
Interquartile range (IQR): Outliers can be identified by calculating the interquartile range (IQR), which is the difference between the first and third quartile. Values that fall outside of 1.5 * IQR above the third quartile or 1.5 * IQR below the first quartile are considered outliers.
- Imputing the missing values: This method can be used to replace outliers with the mean, median, or mode of the dataset.
Feature transformation is the process of applying mathematical functions to the features in a dataset in order to change their distribution or to extract additional information from the data. Feature transformations are often used to improve the performance of machine learning models by making the features more suitable for the model.
-
Label encoding is a method of converting categorical variables, represented as text values, into numerical values. It assigns a unique numerical value to each category or level of a categorical feature. This is often used as a preprocessing step before training a machine learning model.
-
Target encoding is a technique used in machine learning to encode categorical variables. It replaces each category with the average value of the target variable for that category. This can improve the performance of a model by allowing it to better handle categorical variables. It is also known as mean encoding or probability encoding.
Feature scaling is the process of normalizing the range of values for each feature in a dataset. This is often done to ensure that all features are on a similar scale and to prevent some features from having a greater impact on the outcome of a machine learning model than others.
- StandardScaler is a pre-processing method in machine learning used to standardize a dataset by subtracting the mean and scaling to unit variance. It is commonly used for feature scaling before applying a supervised learning algorithm to a dataset.
-
Logistic regression is a statistical method used for predicting binary outcomes (i.e. outcomes with two possible results, such as success or failure). It is a type of generalized linear model (GLM) that is used to model a binary dependent variable based on one or more independent variables.
-
Naive Bayes classifier is a probabilistic machine learning algorithm that is based on the Bayes' theorem, which states that the probability of a hypothesis (in this case, a class label) given some observed evidence (in this case, a feature vector) is equal to the probability of the evidence given the hypothesis, multiplied by the prior probability of the hypothesis, divided by the overall probability of the evidence.
-
K-Nearest Neighbors (KNN) is a type of instance-based, or lazy, learning algorithm. It is a classification algorithm that is used to assign a class label to an unlabeled observation based on the class labels of the k-nearest observations to it in feature space.
-
A Decision Tree Classifier is a type of supervised learning algorithm (having a pre-defined target variable) that is mostly used in classification problems. It works by recursively partitioning the dataset into subsets based on the values of the input features.
-
Support Vector Classifier (SVC) is a type of supervised learning algorithm that can be used for classification and regression tasks. The main idea behind SVC is to find the best hyperplane (a decision boundary) that separates the different classes in the feature space. The best hyperplane is the one that maximizes the margin, which is the distance between the hyperplane and the closest data points from each class, also known as support vectors.
- Confusion Matrix
A confusion matrix is a table that is used to define the performance of a classification algorithm. Each row of the matrix represents the instances in a predicted class, while each column represents the instances in an actual class (or vice versa). The name stems from the fact that it makes it easy to see if the system is confusing two classes (i.e. commonly mislabeling one as another). The diagonal elements represent the number of points for which the predicted label is equal to the true label, while off-diagonal elements are those that are mislabeled by the classifier. It is a useful tool for understanding the performance of a classification algorithm, including the types of errors that the classifier is making.
Confusion Matrix accuracy of Algorithms:
- Logistic regression: 0.8053435114503816
- Naive Bayes classifier: 0.7709923664122137
- K-Nearest Neighbors (KNN): 0.7366412213740458
- A Decision Tree Classifier: 0.7404580152671756
- Support Vector Classifier (SVC): 0.7977099236641222
We chose the Logistic regression because the highest Confusion matrix accuracy.
Hyperparameter tuning is the process of selecting the best set of hyperparameters for a machine learning model. Hyperparameters are parameters that are not learned from the data, but are set before training the model.
Grid search: is a technique used to tune the hyperparameters of a machine learning model. It is a systematic way of going through multiple combinations of parameter settings, cross-validating as it goes, and returning the best set of parameters that yield the highest performance for a given model. The technique involves specifying a set of values for each hyperparameter, creating a "grid" of all possible combinations of those values, and then training and evaluating a model for each combination of values.
Hyperparameters for Logistic regression :
- solvers: ['newton-cg', 'lbfgs', 'liblinear']
- penalty: ['l2']
- c_values: [100, 10, 1.0, 0.1, 0.01]
Confusion matrix accuracy for Logistic regression After Hyper Tuning(Complete dataset): 0.7760416666666666
So, here we are getting better results with Logistic regression. Hence we are selecting a Logistic regression as the main model for the web app.
Flask is a web framework for building web applications using the Python programming language. It is a micro-framework that provides the basic functionality needed to build web applications, such as routing and request handling, without including a lot of additional features or libraries. This makes it lightweight and easy to use, but also allows for flexibility and customization.
After inserting values for the attribute such as the Glucose, age, and Blood Pressure our flask web app will predict the whether report indicates the presence of diabetes or not.
To download a diabetes detection, you can use the git clone command. This command creates a copy of the repository in a new directory on your local machine.
git clone https://github.com/nileshparab42/Diabetes-Detection.git
To set up the project, you can use the pip command to install the required packages specified in the requirements.txt file.
pip install -r requirements.txt
This will install all of the required packages for the project. If you are using a virtual environment for the project, you should activate the environment before running this command.
pip install virtualenv
virtualenv myenv
source myenv/bin/activate
pip install -r requirements.txt
This will create a new virtual environment called myenv, activate it, and then install the required packages.
You can also specify the notebook file location or the working directory, after the command "jupyter notebook" or "jupyter lab" like:
jupyter notebook /path/to/notebook/directory
or
jupyter lab /path/to/notebook/directory
It will open the Jupyter notebook or lab in the specified directory, making it easier to navigate to the notebook you want to open.
To install a Flask project, you can use the following commands in your command line interface:
Create a new directory for your project and navigate to it in the command line:
mkdir myproject
cd myproject
Create and activate a virtual environment using virtualenv:
virtualenv venv
source venv/bin/activate
or conda:
conda create --name myenv
conda activate myenv
Install Flask and any other required dependencies by running:
pip install flask
and
pip install -r requirements.txt
the project has a requirements file.
Create a new file named app.py or main.py in the project directory:
Copy code
touch app.py
In the app.py or main.py file, import Flask, create an instance of the Flask class, and define the routes and functions that will handle the requests.
Run the application by executing:
flask run
or
python app.py
or
python main.py
This will start the development server and make the application accessible at "http://localhost:5000"
If you want to deploy your application to a production environment, you can use a web server like Gunicorn or uWSGI to serve the application. It's worth noting that you can also use frameworks like Flask-CLI or Flask-Script to create more advanced command-line scripts for your application.
- Nilesh Parab (Project Lead) - Website
- This project was inspired by the work of the CodeWithHarry.
- We also used resources and tools from the GeeksforGeeks to develop and test our project.