Compare Regression Algorithms

[github-actions[bot]]

size-limit report 📦

Path	Size
dist/perry.cjs.production.min.js	311.67 KB (-0.2% 🔽)
dist/perry.esm.js	309.35 KB (-0.4% 🔽)

[qur786]

Multivariate Regression algorithm

1. when options: { intercepts: false }

   Execution Time:  35 ms
   Root Mean Square Error:0.6812213296719356
   Success Percentage:62.96874999999999

2. when options: { statistics: false}

   Execution Time:  33 ms
   Root Mean Square Error:0.6766184301362179
   Success Percentage:62.65625

3. when options: { statistics: false, intercepts: false }

   Execution Time:  27 ms
   Root Mean Square Error:0.6812213296719356
   Success Percentage:62.96874999999999

4. when options: { statistics: true, intercepts: true}

   Execution Time:  69 ms
   Root Mean Square Error:0.6766184301362179
   Success Percentage:62.65625

By default intercept and statistics in the Multivariate Regression algorithm are true, so { intercept: false } is equivalent to { intercept: false, statistics: true}

[qur786]

The results are optimum when the test size is 95%

MultivariateLinearRegression
Execution Time:  41 ms
Root Mean Square Error:0.6020797289396148
Success Percentage:67.5
RandomForestRegressions
Execution Time:  90072 ms
Root Mean Square Error:0.75
Success Percentage:51.24999999999999

[qur786]

@maneetgoyal sir using other options is not improving the Random Forest Algorithm model

[qur786]

@maneetgoyal sir there is 150 possible combinations of hyper Perameter with the options object created in the code for Random Forest Algorithm, 1 processing of `Random forest is taking almost 80-90 sec , should we test various predictions based upon the hyperPerameter combinations created with kombi

[maneetgoyal]

hyper-par optimi. to be done via https://github.com/manufac-analytics/perry/issues/84