A data science team is building a model to predict upcoming weather patterns using a dataset with hundreds of climate-related features, including temperature, humidity, wind speed, and atmospheric pressure collected over time. They need a model that can handle high-dimensional data and capture complex, non-linear relationships between variables. Which model would be the most appropriate for this task?

Question

BlackTom AI · Accepted Answer

When tackling a task that involves predicting weather patterns from a high-dimensional, time-series-like climate dataset with hundreds of features, we need a model that can handle many input variables and capture complex, non-linear relationships.

Option 1: Linear Regression. While it is simple and fast, linear regression assumes a linear relationship between features and the target. Weather systems are typically governed by nonlinear interactions among variables (e.g., humidity, temperature, pressure), so linear regression would likely underfit and miss important patterns.

Option 2: Support Vector Machine (SVM). SVMs, especially with a non-linear kernel such as the radial basis function (RBF) or polynomial kernel, are well-suited for high-dimensional spaces and can model intricate relationships between features. They work by finding a decision surface (or regression function) in a transformed feature space that captures non-linear patterns. This makes SVM a strong candidate for complex, high-dimensional data where the relationships are not simply linear. However, one caveat is that SVMs can be computationally intensive on very large datasets, so in practice one might balance this with kernel choice and data sampling.

Option 3: K-Nearest Neighbors (KNN). KNN is simple and non-parametric, but it struggles in high-dimensional spaces due to the curse of dimensionality: distances become less informative, leading to poor generalization and scalability issues with many features. It also lacks a built-in mechanism to model complex interactions beyond local neighborhoods, making it less ideal for this task.

Option 4: Naïve Bayes. This classifier assumes feature independence given the class, which is rarely true in climate data where many features are interdependent (e.g., temperature and humidity, pressure and wind). Its strong independence assumption makes it ill-suited for capturing the nuanced, joint distributions needed to predict weather patterns, especially in high dimensions.

In summary, linear regression falls short due to nonlinearity; KNN and Naïve Bayes have fundamental limitations in high-dimensional, interdependent climate data; SVM, particularly with a non-linear kernel, is designed to handle high-dimensional feature spaces and non-linear relationships, aligning best with the described modeling goal. "

IS 4487-006 Fall 2025 Week 12 - Comprehension Quiz

查看解析

登录即可查看完整答案

类似问题

An SVM is chosen for a clearly separable dataset. What does SVM aim to optimize?

Support Vector Machines aim to maximise the margin between classes

Support Vector Machines aim to maximise the margin between classes

All else being equal, the decision boundary of an SVM model gets smoother (simpler; refer to the image below) as the regularization parameter (a.k.a. a cost parameter) increases.

In the following featured diagram, which line is most likely to be selected by SVM as the decision boundary that classifies the two classes (crosses vs. dots)?

Which of the following correctly describes Support-Vector Machines (SVMs)?

Which of the following satisfies the support vectors? Suppose the linear model, h, is known, and the true label y is in {−1,1}.

更多留学生实用工具

考试浏览器助手

风格化写作助手

论文查重助手

文献引用助手

课堂转译助手

课堂笔记助手

Quiz搜索助手

学校历年真题

智能刷题助手

智能匹配练习

希望你的学习变得更简单