Regression is (along with many different issues) a supervised machine studying methodology, geared toward predicting continuous-valued outputs primarily based on a set of numerical inputs (sometimes called options). It finds plenty of use instances in actual life starting from gross sales forecasting and inventory market prediction to upkeep prediction for machines in a manufacturing or a producing setting.
In a regression evaluation, the mannequin skilled and developed is evaluated utilizing 3 error phrases: SSR, SSE and SST (definitions to comply with). Mathematically, the SST denotes the overall variability within the mannequin efficiency and is expressed because the sum of SSR (the variability defined by the mannequin) and SSE (the unexplained variability). When exploring mathematically why this holds good, I discovered an intriguing method to show this relationship which I needed to share with you.
I outline these 3 error phrases and their mathematical equation on this article. Then, I present the step-by-step process to derive the connection mentioned above. Let’s go!
