In this repository, I am sharing a mini project I did on Statistical Methods on a dataset comprising of 699 breast tissue samples from 699 inviduals from Wisconsin, USA.
Breast Cancer Classification using Statistical Techniques
- R for the data cleaning and analysis.
- Logistic Regression for classification into two classes (Begnin and Malignant)
Non-technical people and the decision makers who may not have statistical knowledge, they tend to focus more on the numbers and that's what I tried to focus throughout my work. To illustrate work, I tried to use the statistical methods to showcase those numbers boldly.
EVALUATIONS: The implementation of logistic regression, LASSO, and Linear Discriminant Analysis (LDA) models offers insights into effective classifiers for distinguishing between benign and malignant tumors using cytological predictors. Logistic regression analysis identifies crucial predictors linked to malignancy, while LASSO highlights influential variables by shrinking some coefficients toward zero. Evaluating these models alongside the robustness of LDA underscores the importance of feature selection and model comparison in developing accurate and efficient diagnostic tools for tumor classification using the current dataset. Additionally, the understanding gained from these models aids in identifying key cytological features pivotal in distinguishing tumor types and refining classification accuracy. The significance of these analysis lies in their role in identifying pivotal cytological features essential for accurate tumor classification with the existing dataset. By leveraging logistic regression, LASSO, and LDA models, this analysis contributes to the development of effective diagnostic tools crucial for enhancing clas- sification accuracy in distinguishing between benign and malignant tumors.
Analysis Report on Newcastle University's Online Cyber Security Course on FutureLearn
- R for the data cleaning and analysis.
EVALUATIONS: In my second cycle of CRISP-DM, while I've gained valuable insights into the correlation between experience ratings and feedback, the constraints of limited data restricted the use of more advanced analytical techniques. I relied on my instincts to assess sentiment, endeavoring to ensure the accuracy of the derived sentiments. Nevertheless, the findings present actionable opportunities for program enhancement. They highlight the significance of refining feedback collection methods and comprehending participants' sentiments to enhance the online learning experience. Here are four key takeaways from my analysis that would potentially answer the question I had put forward.
- Live Demo: KVPS Analysis Dashboard
- Purpose: Explore an open dataset using interactive filters and visualizations
- Tools Used: R, Shiny, Public Policy Dataset
Highlights:
- Interactive data exploration for users of all skill levels
- Clean visualizations to convey insights clearly
- Demonstrates how raw data can be structured for decision-making
Dashboard Preview (Placeholders for GitHub images)
> This project provides a practical example of turning a dataset into actionable insights.