Press train -------------- Then set features and labels as your wish In layman terms , features : your input ,labels : your output --------------- Then press train to train the model It will ask learning_rate and epochs In layman terms, epoch : times of loop for accuracy , learning_rate : how big the jump of variables are there in each epoch/loop ---------------- It will take some time in training Then predict button will show up to predict new data(labels) ---------------- Some info Keeping learning rate very high like 1 could make model inaccurate while keeping it too low like 0.00001 could need more epochs(which will need more time to process) to get to lowest error . ---------------. Enjoy !! --------------- Huge update : After training ,graph is shown with your data and the predicted line !
Could I say Udemy !? Well at first it was random idea to make machine learning in Scratch but hey it works ! Sample Data (50 rows , w=3,x:10 to 50,y:50 to 150:noise:+-1) features : 10.3359 ,11.3942 ,12.3184 ,11.4680,12.4406 ,15.0270 ,14.6609,16.4552 ,16.9622 ,16.9215 ,18.7715 ,19.0977 ,20.4355 ,20.9429 ,21.0758 ,22.4566 ,22.5925 ,23.1349 ,24.1767 ,25.0549 ,26.8888 ,27.6637 ,28.7774 ,28.7472 ,29.5714 ,31.0037 ,31.4027 ,31.7049 ,32.0292 ,33.7364 ,33.8153 ,36.0644 ,36.9397 ,37.7299 ,37.2089 ,38.1103 ,38.8337 ,39.7934 ,40.2194 ,42.2608 ,41.8265,43.4512 ,43.8084 ,44.3005 ,46.1721 ,46.8344 ,46.6844 ,47.5933 ,49.7117 ,49.8639 labels : 50.6629 ,51.6489 ,54.9497 ,56.2395 ,58.0951 ,59.9522 ,61.8572 ,64.1628 ,66.1974 ,67.8382 ,70.3429 ,72.6531 ,74.4854 ,76.8976 ,77.7676 ,80.5700 ,81.8334 ,83.8654 ,86.6385 ,88.1086 ,90.6868 ,93.1168 ,95.6196 ,96.0758 ,98.0592 ,100.6241 ,103.1536 ,104.2392 ,108.0229 ,108.6896 ,111.5550 ,113.0936 ,114.8504 ,116.8315 ,118.5901 ,122.3996 ,122.71387 ,125.2375 ,127.1371 ,128.9257 ,131.1648 ,132.8267 ,134.9273 ,137.2888 ,139.4648 ,142.7921 ,143.9985 ,146.8264 ,148.7271 ,150.3676 recommended : learning rate=0.001 , epochs= 3000 .
