Scribd
Upload
15 views

Training Seminar

Uploaded by

Rishi Chourasia
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
15 views

Training Seminar

Uploaded by

Rishi Chourasia
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 12

INDIAN INSTITUTE OF TECHNOLOGY ROORKEE

Training Seminar Presentation

Weather Data Analysis

By: Mohit Kumar Maan


21112071
Contents :-

• Problem Statement​
• Objectives​
• Data Overview​
• Machine Learning Techniques & Approach​
• Key Insights & Visualizations​
• Conclusion​
• Internship Certificate​

2
Problem statement:-

• Difficulty in accurate weather and air quality prediction using


traditional statistical models.
• Need for advanced methods to handle environmental data’s high
variability and non-linear patterns.
• Essential to improve environmental predictions for sectors like
health, agriculture, and disaster management.
• Demand for reliable machine learning models tailored to handle
multidimensional weather data.

3
Objective :-

• Develop and evaluate machine learning models for accurate


environmental data classification.
• Use feature engineering and preprocessing for improved
model reliability.
• Compare various algorithms for insights into prediction
accuracy and interpretability.

4
Data Overview:-

• Dataset includes weather variables such as temperature,


humidity, wind speed, and pollutant levels.
• Data sourced from Open-Meteo API, ensuring wide
geographical and temporal coverage.
• Preprocessed for quality, handling missing values and
standardizing features.

5
MACHINE LEARNING TECHNIQUES AND APPROACH​

Algorithms Used:​
• KNN: For simplicity and effectiveness.
• Decision Tree: For interpretable , rule-based classification of data

Steps:​
• Data Cleaning: Handled missing values, outliers, remove
duplicates, standardized data.​
• Feature Engineering: Created new features based on customer
behaviour.​
• Model Training: Applied KNN and Decision Tree for predicting
weather.

6
KNN:-
• KNN classifies data points based on the majority class of their neighbors.

• It uses distance measures, like Euclidean, to find nearest neighbors.

• The parameter K determines how many neighbors to consider for classification.

• KNN does not assume any underlying data distribution, making it flexible.

7
Decision Tree

• Decision Tree is a supervised ML algorithm that splits data


into branches based on feature values, creating a tree-like
structure.
• Each node represents a feature decision, leading to
branches for each possible outcome.
• Follows paths from root to leaf, where each leaf represents a
classification or prediction.

8
Key Insights :-

• Decision Tree achieve highest accuracy of 81.54.


• SHAP and LIME analyses highlight feature contributions.
• Understand feature dependencies for model accuracy.
• Compare metrics to highlight model strengths.

9
Conclusion:-

• This project shows the effectiveness of machine learning


models, especially KNN and Decision tree having accuracy
of 76.67 and 81.54 respectively, in accurately classifying and
predicting environmental data. By combining data
preprocessing, model tuning, and feature analysis, it offers a
strong framework for reliable weather and air quality
predictions, with applications in public health and
environmental policy. Future work could explore deeper
learning techniques and broader datasets.

10
Internship Certificate :-

11
Thank You

12

You might also like