- 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
- Hands-On Internet of Things with MQTT
- 電力自動化實用技術問答
- Mastering Elastic Stack
- 新手學電腦快速入門
- Storm應用實踐:實時事務處理之策略
- 數據掘金
- 工業機器人運動仿真編程實踐:基于Android和OpenGL
- PVCBOT機器人控制技術入門
- 嵌入式操作系統原理及應用
- Word 2007,Excel 2007辦公應用融會貫通
- IBM? SmartCloud? Essentials
- 基于RPA技術財務機器人的應用與研究
- WOW!Photoshop CS6完全自學寶典
- AVR單片機工程師是怎樣煉成的
- Effective Business Intelligence with QuickSight