- Deep Learning By Example
- Ahmed Menshawy
- 212字
- 2021-06-24 18:52:40
Motivation
Linear regression models are the building blocks of many learning algorithms, but this is not the only reason behind their popularity. The following are the key factors behind their popularity:
- Widely used: Linear regression is the oldest regression technique and it's widely used in many applications, such as forecasting and financial analysis.
- Runs fast: Linear regression algorithms are very simple and don't include mathematical computations which are too expensive.
- Easy to use (not a lot of tuning required): Linear regression is very easy to use, and mostly it's the first learning method to learn about in the machine learning or data science class as you don't have too many hyperparameters to tune in order to get better performance.
- Highly interpretable: Because of its simplicity and ease of inspecting the contribution of each predictor-coefficient pair, linear regression is highly interpretable; you can easily understand the model behavior and interpret the model output for non-technical guys. If a coefficient is zero, the associated predictor variable contributes nothing. If a coefficient is not zero, the contribution due to the specific predictor variable can easily be ascertained.
- Basis for many other methods: Linear regression is considered the underlying foundation for many learning methods, such as neural networks and its growing part, deep learning.
推薦閱讀
- 智能傳感器技術與應用
- 物聯網與云計算
- INSTANT Varnish Cache How-to
- 樂高創意機器人教程(中級 下冊 10~16歲) (青少年iCAN+創新創意實踐指導叢書)
- 運動控制系統應用與實踐
- SAP Business Intelligence Quick Start Guide
- 分析力!專業Excel的制作與分析實用法則
- 零起點學西門子S7-200 PLC
- INSTANT Heat Maps in R:How-to
- Mastering Ceph
- 電腦故障排除與維護終極技巧金典
- Oracle 11g Anti-hacker's Cookbook
- 創客機器人實戰:基于Arduino和樹莓派
- TensorFlow 2.0卷積神經網絡實戰
- Cloud Native Development Patterns and Best Practices