PML Assignment Report
David A York
November 7, 2015
Executive Summary
This project assignment is to study a model to control for the effect of proper activity teqhnique by constructing a model to predict the occurence deviation from technique. The data to be use to construct a prediction model is from the Human Activity Recognition. Selected analysis focuses on use of data from accelerometers on the belt, forearm, arm, and dumbell of 6 participants.
The resulting random forest based model was able to predice but the true resukts were blinded. The Accuracy (Error) estimation from the resampling used was 99.3% (error 0.7%). The project repository is accessible at on Github here.
Model Requirements (Purpose)
Devices such as Jawbone Up, Nike FuelBand, and Fitbit are able to collect a large amount of data about personal activity (of individual whereing the devices) as part of the quantified self movement (QSM).
People in the QSM regularly quantify how much of a particular activity they do, but they rarely quantify how well they do it. The effectiveness of such activities is assumed to be related to carrying it out with proper technique.
Data was collected from accelerometers on the belt, forearm, arm, and dumbell of 6 participants asked to perform barbell lifts correctly and incorrectly in 5 different ways. Using this data, a model is build to detect improper activity teqhnique in Human Activity Recognition (HAR) data by constructing a prediction model for the occurence of deviation from correct activity technique.
Data Cleaning
The data to be used to construct the prediction model is from the Human Activity Recognition. This was downloaded in edited form from a location specifically for this assignment (see Appendix II Assignment Criteria as set out by instructors, Leek et al.) Selected analysis focuses on use of data from accelerometers on the belt, forearm, arm, and dumbell of 6 participants. It should be noted that this subset of the datasets has missing data in the var_accel_ data for each body part.
Data Exploration
See the structure on the training data selected for the model in Appendix Table 1. See the paires plot in Appendix Figure 1
Model Building
The model resulting is,
## Random Forest
##
## 19622 samples
## 52 predictor
## 5 classes: 'A', 'B', 'C', 'D', 'E'
##
## No pre-processing
## Resampling: Bootstrapped (25 reps)
## Summary of sample sizes: 19622, 19622, 19622, 19622, 19622, 19622, ...
## Resampling results across tuning parameters:
##
## mtry Accuracy Kappa Accuracy SD Kappa SD
## 2 0.9930377 0.9911913 0.001194638 0.001512108
## 27 0.9928538 0.9909588 0.001092238 0.001382576
## 52 0.9858949 0.9821553 0.003117982 0.003940709
##
## Accuracy was used to select the optimal model using the largest value.
## The final value used for the model was mtry = 2.##
## Call:
## randomForest(x = x, y = y, mtry = param$mtry)
## Type of random forest: classification
## Number of trees: 500
## No. of variables tried at each split: 2
##
## OOB estimate of error rate: 0.44%
## Confusion matrix:
## A B C D E class.error
## A 5577 2 0 0 1 0.0005376344
## B 13 3779 5 0 0 0.0047405847
## C 0 16 3405 1 0 0.0049678551
## D 0 0 40 3174 2 0.0130597015
## E 0 0 0 6 3601 0.0016634322
Model Validation
Model validation is inherent in the resampling of the train() function. The validity is determine through the metric(s) chosen to drive the resampling with the accuracy and error rate visualized below from the confusion matrix and the means of the resampling accuracy and kappa.
Expected Error
The expected error is evident from the model result above. The actual accuracy with the test set cannot be derived as these are the unknowns, being blinded for the Questions 1 to 20. Checked elsewhere.
Predicting on the test data with the model model built with the random forest method obtains,
## X user_name raw_timestamp_part_1 raw_timestamp_part_2 cvtd_timestamp
## 1 1 pedro 1323095002 868349 05/12/2011 14:23
## 2 2 jeremy 1322673067 778725 30/11/2011 17:11
## 3 3 jeremy 1322673075 342967 30/11/2011 17:11
## 4 4 adelmo 1322832789 560311 02/12/2011 13:33
## 5 5 eurico 1322489635 814776 28/11/2011 14:13
## 6 6 jeremy 1322673149 510661 30/11/2011 17:12
## 7 7 jeremy 1322673128 766645 30/11/2011 17:12
## 8 8 jeremy 1322673076 54671 30/11/2011 17:11
## 9 9 carlitos 1323084240 916313 05/12/2011 11:24
## 10 10 charles 1322837822 384285 02/12/2011 14:57
## 11 11 carlitos 1323084277 36553 05/12/2011 11:24
## 12 12 jeremy 1322673101 442731 30/11/2011 17:11
## 13 13 eurico 1322489661 298656 28/11/2011 14:14
## 14 14 jeremy 1322673043 178652 30/11/2011 17:10
## 15 15 jeremy 1322673156 550750 30/11/2011 17:12
## 16 16 eurico 1322489713 706637 28/11/2011 14:15
## 17 17 pedro 1323094971 920315 05/12/2011 14:22
## 18 18 carlitos 1323084285 176314 05/12/2011 11:24
## 19 19 pedro 1323094999 828379 05/12/2011 14:23
## 20 20 eurico 1322489658 106658 28/11/2011 14:14
## new_window num_window roll_belt pitch_belt yaw_belt total_accel_belt
## 1 no 74 123.00 27.00 -4.75 20
## 2 no 431 1.02 4.87 -88.90 4
## 3 no 439 0.87 1.82 -88.50 5
## 4 no 194 125.00 -41.60 162.00 17
## 5 no 235 1.35 3.33 -88.60 3
## 6 no 504 -5.92 1.59 -87.70 4
## 7 no 485 1.20 4.44 -87.30 4
## 8 no 440 0.43 4.15 -88.50 4
## 9 no 323 0.93 6.72 -93.70 4
## 10 no 664 114.00 22.40 -13.10 18
## 11 no 859 0.92 5.94 -92.70 3
## 12 no 461 1.01 4.96 -87.80 5
## 13 no 257 0.54 2.45 -88.60 3
## 14 no 408 0.45 5.02 -87.90 5
## 15 no 779 5.34 -3.09 -80.30 4
## 16 no 302 1.65 3.47 -87.00 2
## 17 no 48 129.00 27.80 1.84 21
## 18 no 361 0.92 5.31 -93.10 3
## 19 no 72 123.00 26.70 -2.68 19
## 20 no 255 1.40 3.20 -88.70 3
## gyros_belt_x gyros_belt_y gyros_belt_z accel_belt_x accel_belt_y
## 1 -0.50 -0.02 -0.46 -38 69
## 2 -0.06 -0.02 -0.07 -13 11
## 3 0.05 0.02 0.03 1 -1
## 4 0.11 0.11 -0.16 46 45
## 5 0.03 0.02 0.00 -8 4
## 6 0.10 0.05 -0.13 -11 -16
## 7 -0.06 0.00 0.00 -14 2
## 8 -0.18 -0.02 -0.03 -10 -2
## 9 0.10 0.00 -0.02 -15 1
## 10 0.14 0.11 -0.16 -25 63
## 11 0.05 0.00 0.00 -18 4
## 12 -0.10 0.00 0.02 -22 8
## 13 -0.06 0.00 -0.13 -8 5
## 14 -0.05 0.00 0.00 -14 2
## 15 0.24 0.00 0.05 8 19
## 16 0.02 0.02 0.00 -12 5
## 17 -0.50 0.00 -0.48 -47 69
## 18 0.02 0.00 -0.05 -13 3
## 19 -0.31 -0.05 -0.44 -48 72
## 20 0.06 -0.02 0.02 -9 4
## accel_belt_z magnet_belt_x magnet_belt_y magnet_belt_z roll_arm
## 1 -179 -13 581 -382 40.70
## 2 39 43 636 -309 0.00
## 3 49 29 631 -312 0.00
## 4 -156 169 608 -304 -109.00
## 5 27 33 566 -418 76.10
## 6 38 31 638 -291 0.00
## 7 35 50 622 -315 0.00
## 8 42 39 635 -305 0.00
## 9 32 -6 600 -302 -137.00
## 10 -158 10 601 -330 -82.40
## 11 27 6 599 -312 70.00
## 12 40 51 632 -308 0.00
## 13 24 34 571 -426 110.00
## 14 49 55 635 -305 0.00
## 15 28 91 584 -398 0.00
## 16 20 45 566 -425 88.10
## 17 -187 4 566 -400 -1.47
## 18 24 0 607 -305 152.00
## 19 -169 -5 584 -366 30.70
## 20 23 37 567 -425 90.60
## pitch_arm yaw_arm total_accel_arm gyros_arm_x gyros_arm_y gyros_arm_z
## 1 -27.80 178.0 10 -1.65 0.48 -0.18
## 2 0.00 0.0 38 -1.17 0.85 -0.43
## 3 0.00 0.0 44 2.10 -1.36 1.13
## 4 55.00 -142.0 25 0.22 -0.51 0.92
## 5 2.76 102.0 29 -1.96 0.79 -0.54
## 6 0.00 0.0 14 0.02 0.05 -0.07
## 7 0.00 0.0 15 2.36 -1.01 0.89
## 8 0.00 0.0 22 -3.71 1.85 -0.69
## 9 11.20 -167.0 34 0.03 -0.02 -0.02
## 10 -63.80 -75.3 32 0.26 -0.50 0.79
## 11 15.10 113.0 33 -0.08 0.02 0.00
## 12 0.00 0.0 30 2.57 -1.06 0.79
## 13 -31.70 -13.6 23 -3.44 1.08 -0.15
## 14 0.00 0.0 37 -0.02 -0.02 -0.10
## 15 0.00 0.0 23 3.66 -2.09 0.49
## 16 5.58 123.0 33 0.02 -0.06 0.10
## 17 -3.45 -55.1 3 -0.19 -0.24 0.13
## 18 -26.40 -85.0 42 1.00 -0.37 -0.03
## 19 -27.00 -151.0 11 1.99 -1.20 -0.44
## 20 11.50 117.0 30 -0.47 0.13 -0.18
## accel_arm_x accel_arm_y accel_arm_z magnet_arm_x magnet_arm_y
## 1 16 38 93 -326 385
## 2 -290 215 -90 -325 447
## 3 -341 245 -87 -264 474
## 4 -238 -57 6 -173 257
## 5 -197 200 -30 -170 275
## 6 -26 130 -19 396 176
## 7 99 79 -67 702 15
## 8 -98 175 -78 535 215
## 9 -287 111 -122 -367 335
## 10 -301 -42 -80 -420 294
## 11 -277 113 -112 -320 350
## 12 -192 204 -75 194 280
## 13 106 97 -168 697 -184
## 14 -277 157 -183 -404 416
## 15 41 85 -204 750 -24
## 16 -236 163 -148 -398 288
## 17 2 -9 27 -394 321
## 18 35 -65 -404 202 -307
## 19 -7 57 89 -266 454
## 20 -223 166 -105 -428 321
## magnet_arm_z roll_dumbbell pitch_dumbbell yaw_dumbbell
## 1 481 -17.737480 24.96085 126.235964
## 2 434 54.477605 -53.69758 -75.514799
## 3 413 57.070308 -51.37303 -75.202873
## 4 633 43.109268 -30.04885 -103.320032
## 5 617 -101.383964 -53.43952 -14.195423
## 6 516 62.187498 -50.55595 -71.120631
## 7 217 56.046184 -53.96761 -73.870552
## 8 385 59.128080 -51.30242 -73.368487
## 9 520 122.524075 0.00000 34.555558
## 10 493 -111.117700 -31.90635 31.906354
## 11 514 123.984144 -32.07139 7.889285
## 12 555 45.507300 -53.70901 -83.050116
## 13 -175 12.479968 34.27571 120.424746
## 14 409 56.370775 -49.97065 -77.029188
## 15 -58 46.327501 -54.97284 -81.274990
## 16 525 57.795891 33.72223 87.799283
## 17 502 -28.497111 13.61379 125.331471
## 18 -499 107.472223 -49.53990 -2.377137
## 19 441 -7.463022 23.65339 132.233740
## 20 472 36.916678 96.86773 45.178263
## total_accel_dumbbell gyros_dumbbell_x gyros_dumbbell_y gyros_dumbbell_z
## 1 9 0.64 0.06 -0.61
## 2 31 0.34 0.05 -0.71
## 3 29 0.39 0.14 -0.34
## 4 18 0.10 -0.02 0.05
## 5 4 0.29 -0.47 -0.46
## 6 29 -0.59 0.80 1.10
## 7 29 0.34 0.16 -0.23
## 8 29 0.37 0.14 -0.39
## 9 3 0.03 -0.21 -0.21
## 10 2 0.42 0.51 -0.03
## 11 1 0.18 -0.47 -0.23
## 12 30 0.39 -0.10 -0.56
## 13 8 1.06 -1.11 -1.00
## 14 30 0.35 0.11 -0.38
## 15 31 0.75 -0.21 -1.18
## 16 13 -1.03 0.92 0.13
## 17 11 0.50 -0.11 -0.33
## 18 4 -0.39 1.91 0.52
## 19 10 0.45 -0.10 -0.23
## 20 23 0.79 -0.79 -0.23
## accel_dumbbell_x accel_dumbbell_y accel_dumbbell_z magnet_dumbbell_x
## 1 21 -15 81 523
## 2 -153 155 -205 -502
## 3 -141 155 -196 -506
## 4 -51 72 -148 -576
## 5 -18 -30 -5 -424
## 6 -138 166 -186 -543
## 7 -145 150 -190 -484
## 8 -140 159 -191 -515
## 9 0 25 9 -519
## 10 -7 -20 7 -531
## 11 -4 12 1 -556
## 12 -149 128 -215 -539
## 13 27 10 75 -328
## 14 -139 155 -202 -523
## 15 -159 136 -221 -511
## 16 43 71 100 -284
## 17 14 -29 98 484
## 18 -20 37 -1 -527
## 19 22 -7 91 505
## 20 185 81 98 272
## magnet_dumbbell_y magnet_dumbbell_z roll_forearm pitch_forearm
## 1 -528 -56 141.0 49.30
## 2 388 -36 109.0 -17.60
## 3 349 41 131.0 -32.60
## 4 238 53 0.0 0.00
## 5 252 312 -176.0 -2.16
## 6 262 96 150.0 1.46
## 7 354 97 155.0 34.50
## 8 350 53 -161.0 43.60
## 9 348 -32 15.5 -63.50
## 10 321 -164 13.2 19.40
## 11 280 -23 137.0 -9.41
## 12 293 67 138.0 -19.90
## 13 271 363 176.0 16.70
## 14 337 46 53.1 -46.20
## 15 323 85 152.0 26.50
## 16 311 368 -176.0 16.20
## 17 -558 -19 79.4 -2.60
## 18 329 -77 -173.0 49.80
## 19 -537 -86 -164.0 59.30
## 20 403 340 173.0 19.20
## yaw_forearm total_accel_forearm gyros_forearm_x gyros_forearm_y
## 1 156.0 33 0.74 -3.34
## 2 106.0 39 1.12 -2.78
## 3 93.0 34 0.18 -0.79
## 4 0.0 43 1.38 0.69
## 5 -47.9 24 -0.75 3.10
## 6 89.7 43 -0.88 4.26
## 7 152.0 32 -0.53 1.80
## 8 -89.5 47 0.63 -0.74
## 9 -139.0 36 0.03 0.02
## 10 -105.0 24 0.02 0.13
## 11 104.0 46 0.05 0.05
## 12 70.5 36 -0.02 0.83
## 13 -38.5 23 0.00 -4.69
## 14 159.0 33 0.02 0.00
## 15 -168.0 24 0.64 -5.97
## 16 -42.2 25 -0.31 -3.45
## 17 109.0 30 0.10 -0.06
## 18 -133.0 25 -0.75 4.14
## 19 -149.0 23 -1.06 2.79
## 20 -83.2 21 -1.01 3.18
## gyros_forearm_z accel_forearm_x accel_forearm_y accel_forearm_z
## 1 -0.59 -110 267 -149
## 2 -0.18 212 297 -118
## 3 0.28 154 271 -129
## 4 1.80 -92 406 -39
## 5 0.80 131 -93 172
## 6 1.35 230 322 -144
## 7 0.75 -192 170 -175
## 8 0.49 -151 -331 -282
## 9 -0.02 195 204 -217
## 10 -0.07 -212 98 -7
## 11 0.11 -3 405 -203
## 12 0.38 182 263 -148
## 13 -1.26 149 46 167
## 14 0.03 232 106 -198
## 15 -1.05 -18 43 -226
## 16 -0.66 161 -61 171
## 17 0.18 144 201 -154
## 18 0.95 -129 -29 -202
## 19 0.67 -148 21 -172
## 20 1.26 41 -100 179
## magnet_forearm_x magnet_forearm_y magnet_forearm_z predRight
## 1 -714 419 617 B
## 2 -237 791 873 A
## 3 -51 698 783 B
## 4 -233 783 521 A
## 5 375 -787 91 A
## 6 -300 800 884 E
## 7 -678 284 585 D
## 8 -109 -619 -32 B
## 9 0 652 469 A
## 10 -403 723 512 A
## 11 -248 720 438 B
## 12 32 648 702 C
## 13 456 -677 29 B
## 14 123 555 648 A
## 15 -540 -155 389 E
## 16 532 -775 114 E
## 17 -146 756 708 A
## 18 -500 -232 329 B
## 19 -614 -46 471 B
## 20 70 -703 74 BError Estimation
Next we will review the the accurate and associated miss-classification errors. Consider the confusion matrix.
## Bootstrapped (25 reps) Confusion Matrix
##
## (entries are percentages of table totals)
##
## Reference
## Prediction A B C D E
## A 28.4 0.1 0.0 0.0 0.0
## B 0.0 19.1 0.1 0.0 0.0
## C 0.0 0.0 17.3 0.3 0.0
## D 0.0 0.0 0.0 16.0 0.0
## E 0.0 0.0 0.0 0.0 18.5An estimate of the error rate is evident from the summary in the confusion matrix showing 0.1% of B and C and 0.3% of D were miss-classified, so the error rate is 0.6962256
Appendices
Apendix I: Report Figures
Table 1 The Structure of Actual Training Data for the Model
## 'data.frame': 19622 obs. of 60 variables:
## $ X : int 1 2 3 4 5 6 7 8 9 10 ...
## $ user_name : Factor w/ 6 levels "adelmo","carlitos",..: 2 2 2 2 2 2 2 2 2 2 ...
## $ raw_timestamp_part_1: int 1323084231 1323084231 1323084231 1323084232 1323084232 1323084232 1323084232 1323084232 1323084232 1323084232 ...
## $ raw_timestamp_part_2: int 788290 808298 820366 120339 196328 304277 368296 440390 484323 484434 ...
## $ cvtd_timestamp : POSIXlt, format: "2011-12-05 11:23:00" "2011-12-05 11:23:00" ...
## $ new_window : Factor w/ 2 levels "no","yes": 1 1 1 1 1 1 1 1 1 1 ...
## $ num_window : num 11 11 11 12 12 12 12 12 12 12 ...
## $ roll_belt : num 1.41 1.41 1.42 1.48 1.48 1.45 1.42 1.42 1.43 1.45 ...
## $ pitch_belt : num 8.07 8.07 8.07 8.05 8.07 8.06 8.09 8.13 8.16 8.17 ...
## $ yaw_belt : num -94.4 -94.4 -94.4 -94.4 -94.4 -94.4 -94.4 -94.4 -94.4 -94.4 ...
## $ total_accel_belt : num 3 3 3 3 3 3 3 3 3 3 ...
## $ gyros_belt_x : num 0 0.02 0 0.02 0.02 0.02 0.02 0.02 0.02 0.03 ...
## $ gyros_belt_y : num 0 0 0 0 0.02 0 0 0 0 0 ...
## $ gyros_belt_z : num -0.02 -0.02 -0.02 -0.03 -0.02 -0.02 -0.02 -0.02 -0.02 0 ...
## $ accel_belt_x : num -21 -22 -20 -22 -21 -21 -22 -22 -20 -21 ...
## $ accel_belt_y : num 4 4 5 3 2 4 3 4 2 4 ...
## $ accel_belt_z : num 22 22 23 21 24 21 21 21 24 22 ...
## $ magnet_belt_x : num -3 -7 -2 -6 -6 0 -4 -2 1 -3 ...
## $ magnet_belt_y : num 599 608 600 604 600 603 599 603 602 609 ...
## $ magnet_belt_z : num -313 -311 -305 -310 -302 -312 -311 -313 -312 -308 ...
## $ roll_arm : num -128 -128 -128 -128 -128 -128 -128 -128 -128 -128 ...
## $ pitch_arm : num 22.5 22.5 22.5 22.1 22.1 22 21.9 21.8 21.7 21.6 ...
## $ yaw_arm : num -161 -161 -161 -161 -161 -161 -161 -161 -161 -161 ...
## $ total_accel_arm : num 34 34 34 34 34 34 34 34 34 34 ...
## $ gyros_arm_x : num 0 0.02 0.02 0.02 0 0.02 0 0.02 0.02 0.02 ...
## $ gyros_arm_y : num 0 -0.02 -0.02 -0.03 -0.03 -0.03 -0.03 -0.02 -0.03 -0.03 ...
## $ gyros_arm_z : num -0.02 -0.02 -0.02 0.02 0 0 0 0 -0.02 -0.02 ...
## $ accel_arm_x : num -288 -290 -289 -289 -289 -289 -289 -289 -288 -288 ...
## $ accel_arm_y : num 109 110 110 111 111 111 111 111 109 110 ...
## $ accel_arm_z : num -123 -125 -126 -123 -123 -122 -125 -124 -122 -124 ...
## $ magnet_arm_x : num -368 -369 -368 -372 -374 -369 -373 -372 -369 -376 ...
## $ magnet_arm_y : num 337 337 344 344 337 342 336 338 341 334 ...
## $ magnet_arm_z : num 516 513 513 512 506 513 509 510 518 516 ...
## $ roll_dumbbell : num 13.1 13.1 12.9 13.4 13.4 ...
## $ pitch_dumbbell : num -70.5 -70.6 -70.3 -70.4 -70.4 ...
## $ yaw_dumbbell : num -84.9 -84.7 -85.1 -84.9 -84.9 ...
## $ total_accel_dumbbell: num 37 37 37 37 37 37 37 37 37 37 ...
## $ gyros_dumbbell_x : num 0 0 0 0 0 0 0 0 0 0 ...
## $ gyros_dumbbell_y : num -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 ...
## $ gyros_dumbbell_z : num 0 0 0 -0.02 0 0 0 0 0 0 ...
## $ accel_dumbbell_x : num -234 -233 -232 -232 -233 -234 -232 -234 -232 -235 ...
## $ accel_dumbbell_y : num 47 47 46 48 48 48 47 46 47 48 ...
## $ accel_dumbbell_z : num -271 -269 -270 -269 -270 -269 -270 -272 -269 -270 ...
## $ magnet_dumbbell_x : num -559 -555 -561 -552 -554 -558 -551 -555 -549 -558 ...
## $ magnet_dumbbell_y : num 293 296 298 303 292 294 295 300 292 291 ...
## $ magnet_dumbbell_z : num -65 -64 -63 -60 -68 -66 -70 -74 -65 -69 ...
## $ roll_forearm : num 28.4 28.3 28.3 28.1 28 27.9 27.9 27.8 27.7 27.7 ...
## $ pitch_forearm : num -63.9 -63.9 -63.9 -63.9 -63.9 -63.9 -63.9 -63.8 -63.8 -63.8 ...
## $ yaw_forearm : num -153 -153 -152 -152 -152 -152 -152 -152 -152 -152 ...
## $ total_accel_forearm : num 36 36 36 36 36 36 36 36 36 36 ...
## $ gyros_forearm_x : num 0.03 0.02 0.03 0.02 0.02 0.02 0.02 0.02 0.03 0.02 ...
## $ gyros_forearm_y : num 0 0 -0.02 -0.02 0 -0.02 0 -0.02 0 0 ...
## $ gyros_forearm_z : num -0.02 -0.02 0 0 -0.02 -0.03 -0.02 0 -0.02 -0.02 ...
## $ accel_forearm_x : num 192 192 196 189 189 193 195 193 193 190 ...
## $ accel_forearm_y : num 203 203 204 206 206 203 205 205 204 205 ...
## $ accel_forearm_z : num -215 -216 -213 -214 -214 -215 -215 -213 -214 -215 ...
## $ magnet_forearm_x : num -17 -18 -18 -16 -17 -9 -18 -9 -16 -22 ...
## $ magnet_forearm_y : num 654 661 658 658 655 660 659 660 653 656 ...
## $ magnet_forearm_z : num 476 473 469 469 473 478 470 474 476 473 ...
## $ classe : Factor w/ 5 levels "A","B","C","D",..: 1 1 1 1 1 1 1 1 1 1 ...#Figure 1 Compare the exercises and classe 
Appendix II: Project [R] Code Listing
## R Main Code Block
# figures and tables are call by small code blocks for printing at the appropriate
# places below.
# load required libraries and functions
library(caret)
library(randomForest)
library(ggplot2)
library(GGally)
library(lattice)
library(knitr)
## GET and CLEAN data ##################
# fetch Training and Testing data (if necessary)
if(!file.exists("train.csv")){
fileUrlTrain <- "https://d396qusza40orc.cloudfront.net/predmachlearn/pml-training.csv"
download.file(fileUrlTrain, destfile = "train.csv")
# save the download dates for reference
dateTrainDownloaded <- date()
}
if(!file.exists("test.csv")){
fileUrlTest <-"https://d396qusza40orc.cloudfront.net/predmachlearn/pml-testing.csv"
download.file(fileUrlTest, destfile = "test.csv")
# save the download dates for reference
dateTestDownloaded <- date()
}
## Read in (load) the data files for training and testing ######
training <- read.csv("train.csv")
testing <- read.csv("test.csv")
## end of data aquision ####################
## Clean and tidy data and subsetting #############
# remove and "#DIV/0" to make missing data
divfactorIDX <-training=="#DIV/0!"
training[divfactorIDX] = NA
# cvdt_timestamp, ie var# 5, correct to DATE type (POSIXlt)
# var# 3 and 4 s/b ok as interger (esp. for now)
training$cvtd_timestamp <-strptime(as.character(training$cvtd_timestamp), "%d/%m/%Y %H:%M")
# make variables 7 to 159 all numeric
for(i in c(7:159))
{
training[,i] <-as.numeric(as.character(training[,i]))
}
for(i in c(7:159))
{
testing[,i] <-as.numeric(as.character(testing[,i]))
}
#remove NA columns
# Routine to remove na values (Coursera, PML Discussion, Thread, "Error message when using train(), working on project", 11-22-20156 10:02AM,
# see https://class.coursera.org/predmachlearn-034/forum/thread?thread_id=129&utm_medium=email&u # tm_source=other&utm_campaign=notifications.auto.V_7MppEiEeWqDRKfSaeX2Q#post-484
# by ARJUN KV @ https://github.com/arj27083)
# "You don't get the error if you remove variables with NA values.
# Use the following code to remove it."
# "Note: Must also remove zero variates for which need nearZeroVar() function."
naVar <-function(df){
na_var <- apply(!is.na(df),2,sum) #identifying the no of NAs in each column
na_var <- na_var == nrow(df) #Identifying columns that don't have complete data without NA
df <- df[ , na_var] #Filtering out variables with NA values
return(df)
}
# now clean the NA columns out of the dataframes
training <- naVar(training) # note last training column is now the classe factor variable
testing <- naVar(testing) # note last column is redundant and classe is omitted for testing
testing <- testing[,-60] # remove redundant column
## end of data tidying
## Data exploration #################
## Plots
## scatter plots
plot1 <- qplot(training$classe,training$accel_belt_y, main = "classe vs accel_belt_y")
plot2 <- qplot(training$classe,training$accel_dumbbell_y, main = "classe vs accel_dumbbell_x")
plot3 <- qplot(training$classe,training$accel_arm_z, main = "classe vs accel_arm_z")
plot4 <- qplot(training$classe,training$accel_forearm_z, main = "classe vs accel_forearm_z")
## bar and whisker plots
plot5 <- function(){plot(training$classe,training$accel_belt_x, type="p",main = "classe vs accel_belt_x")}
plot6 <- function(){plot(training$classe,training$accel_dumbbell_y, type="p",main = "classe vs accel_dumbbell_y")}
plot7 <- function(){plot(training$classe,training$accel_arm_z, type="p",main = "classe vs accel_arm_z")}
plot8 <- function(){plot(training$classe,training$accel_forearm_x, type="p",main = "classe vs accel_forearm_x")}
## pairs plot
plot9 <- ggpairs(training[,c(15,29,43,55,60)])
## end data exploration ##
## Build and Assess Model ####################
# built using numeric variables 8:59 as regressors and classe factor varibale (column 60)
# as the response variable
## model validation by the resampling method(s) inherent in the train() function
modelfit52 <- train(classe~.,data = training[,-c(1:7)], method="rf")
# plot groups in data, and their centers
##belt_arm_x <- classCenter(training[,c(15,28)], training$classe, training)
##belt_arm_x <- as.data.frame(belt_arm_x)
##belt_arm_x$classe <- rownames(belt_arm_x)
plot10 <- qplot(accel_belt_x,accel_arm_x, col="classe", data=training)
##plot10 <- plot10+geom_point(aes(accel_belt_x, accel_arm_x,col=classe), size=5, shape=4, data= training)
## Carry out predition using testing data with model fit from the training data
pred <- predict(modelfit52,testing[,-c(1:7)])
testing$predRight <- pred
## Files for 20 questions
answers = as.vector(as.character(pred))
# Function to make files
pml_write_files = function(x){
n = length(x)
for(i in 1:n){
filename = paste0("problem_id_",i,".txt")
write.table(x[i],file=filename,quote=FALSE,row.names=FALSE,col.names=FALSE)
}
}
pml_write_files(answers)
modelfit52
modelfit52$finalModel
print(plot10)
testing
cM <- confusionMatrix(modelfit52)
cM
errRate <- 100 - sum(cM$table[1,1],cM$table[2,2],cM$table[3,3],cM$table[4,4],cM$table[5,5])
# display Training structure
str(training)
# display the pairs plot
plot5()
plot6()
plot9Appendix III Assignment Criteria
Background
Using devices such as Jawbone Up, Nike FuelBand, and Fitbit it is now possible to collect a large amount of data about personal activity relatively inexpensively. These type of devices are part of the quantified self movement – a group of enthusiasts who take measurements about themselves regularly to improve their health, to find patterns in their behavior, or because they are tech geeks. One thing that people regularly do is quantify how much of a particular activity they do, but they rarely quantify how well they do it. In this project, your goal will be to use data from accelerometers on the belt, forearm, arm, and dumbell of 6 participants. They were asked to perform barbell lifts correctly and incorrectly in 5 different ways. More information is available from the HAR website (see the section on the Weight Lifting Exercise Dataset).
Data Source(s)
The data for this project come from the source.
The training data for this project are available for download from https://d396qusza40orc.cloudfront.net/predmachlearn/pml-training.csv
The test data are available for download from https://d396qusza40orc.cloudfront.net/predmachlearn/pml-testing.csv
Purpose
Project goal is to predict the manner in which exercise is performed (technique).
Method
The response is the “classe” variable in the training set. Determine a set of variables which best predicts technique (classe).
The Report describes how the model was built, how cross validation was used, the expected out of sample error, and why the choices were made.
Finally, the prediction model is used to predict 20 different test cases to determine it’s accuracy.
Requirements
- Submit a link to a Github repo with the R markdown and compiled HTML file describing the analysis. The text of the writeup should be < 2000 words and the number of figures less than 5. Submitting a repo with a gh-pages branch is desirable so the HTML page can be viewed online.
- The Machine Learning Algorithm will be applied to the 20 test cases available in the test data above. submitting the predictions, in appropriate format, to the programming assignment for automated grading. (See the programming assignment for additional details.
Submitting the Predictions to the Questions
Apply the machine learning algorithm you built to each of the 20 test cases in the testing data set. See the prediction assignment writeup below on how model to be build. For each test case, submit a text file with a single capital letter (A, B, C, D, or E) corresponding to your prediction for the corresponding problem in the test data set. Score is 1 point for each correct answer. May be submitted up to 2 times for each problem. It is a lot of files to submit so it may be helpful to use the following function to create the files from a character vector with the 20 predictions in correct order for the 20 problems.
Something like (note these are not the right answers!):
answers = rep("A", 20)then load the following function by copying and pasting it into R:
pml_write_files = function(x){
n = length(x)
for(i in 1:n){
filename = paste0("problem_id_",i,".txt")
write.table(x[i],file=filename,quote=FALSE,row.names=FALSE,col.names=FALSE)
}
}create a folder where the files are to be written. Set that to be your working directory and run:
pml_write_files(answers)to create one file for each submission.
Note: using this script, make sure the files that get written out have one character each with the prediction for the corresponding problem ID. The script produces strange results if the answers variable is not a character vector.