Grid Analytics Academy

Happy New Year Wishes-2024.

We have created a youtube video-A father Daughter conversation on How to compete and Win with Data Science-Capstone Projects.
https://youtu.be/_z7ggag8ggU
📢👉Project: Data Science Capstone Project.
✍️Project Title: Prediction of Heart Disease.
💻Methodology: Using CRISP (Cross Industry Standard for Data Mining) framework.
In this example, we will study a classification problem, i.e., a problem with a categorical output (dependent) variable. Our task is to build a Robust classifier model that correctly predicts the class label (category) Coronary Heart Disease label with Present or Absent.
For the Project, we'll use the classic HeartDisease.xls data set. This data set contains information about medical history of 462 of Heart Disease patients with 17 Variables (Column) and 462 cases (Observation).
Dependent (Target) Variable : Coronary Heart Disease (label with 2 classes (Present and Absent).
Independent (Input) Variables: 16 consist of.
Blood Pressure Level,Systolic Blood Pressure,Tobacco Intake Level,Tobacco Intake (kg)
Cholesterol Level,LDL Cholesterol,Adiposity,Family History.,Stress Level Type A,Stress Type-A behaviour,Obesity Level,Obesity,Alcohol Intake Level,Alcohol Consumption,Age Range,Age Disease
📈Data Preparation : We have applied Data Health Check-up , you may see from the link below.
https://lnkd.in/dqxsZ8Fi
Diagnostic Analysis: We applied a Statistica feature selection/Feature Engineering for variable of Importance with p-value and Chi-square computation.
Model Building: We applied all possible classification Model such as Logistic Regression, CART, CHAID, SVM, Bayes and few more.
Evaluation: We evaluated predictive model using various evaluation technique such as: Confusion Matrix, ROC/AUC, Gain/Lift Chart, we found CART was working well with the given data-set, hence we selected CART Model.
Deployment: We generated C # code using Statistica code generator, since we were comfortable with C # coding but not limited.
Statistica can generate code in PMML, C, JAVA, SQL etc… depending on your choice of coding.
We created an executable file using C # for an platform independent API, where API can be deployed in absence of Statistica as a platform independent Model.
If a new patient visit to the clinic,we can predict the status of Coronary Heart Disease using CART model, whether Heart Disease will be present or absent with over 72% confidence probablity accuracy.
On basis of Prediction, we can prescribe the patient the caution and recommendation and prevent the outcome in isntance of present predicted.
Please find below some resources for your self-paced learning.
Statistica™ 30-days Trial download Link: https://lnkd.in/dkEChs4U
Statistica™ How-To-Do Video Tutorials Link: https://lnkd.in/dkEChs4U
# Data Science, # Machine Learning, , Data Science project, # Predictive Analytics # Statistica Statistica, Data Science Analytics.

Please circle this information among students group, peers, network or interest groups for better reach and benefit from this unique Employability/Employment program, your support would be highly appreciated.
We are also looking for internship candidate;
1. Sales/Marketing Executive Intern.
2. Statistician/Data Scientist Intern.
Interested candidate may feel free to contact through e-mail or numbers described below.