turicreate学习笔记1-activity_classif

官方对应文档

1.安装

sudo python -m pip install turicreate

2.activity classifier

数据集下载

• 数据预处理,生成turicreate识别的sFrame格式

from glob import glob
import turicreate as tc

# load labels
labels = tc.SFrame.read_csv('./HAPT Data Set/RawData/labels.txt',
  delimiter=' ',
  header=False,
  verbose=False)
labels = labels.rename({
  'X1': 'exp_id',
  'X2': 'user_id',
  'X3': 'activity_id',
  'X4': 'start',
  'X5': 'end'
  })

def find_label_for_containing_interval(intervals, index):
  containing_interval = intervals[:, 0][(intervals[:, 1] <= index) & (index <= intervals[:, 2])]
  if len(containing_interval) == 1:
    return containing_interval[0]

# load data
data = tc.SFrame()
acc_files = glob('./HAPT Data Set/RawData/acc_*.txt')
gyro_files = glob('./HAPT Data Set/RawData/gyro_*.txt')
files = zip(sorted(acc_files), sorted(gyro_files))
for acc_file, gyro_file in files:
  exp_id = int(acc_file.split('_')[1][-2:])
  user_id = int(acc_file.split('_')[2][4:6])

  # load accel data
  sf = tc.SFrame.read_csv(acc_file,
    delimiter=' ',
    header=False,
    verbose=False)
  sf = sf.rename({
    'X1': 'acc_x',
    'X2': 'acc_y',
    'X3': 'acc_z'
    })
  sf['exp_id'] = exp_id
  sf['user_id'] = user_id

  # load gyro data
  gyro_sf = tc.SFrame.read_csv(gyro_file,
    delimiter=' ',
    header=False,
    verbose=False)
  gyro_sf = gyro_sf.rename({
    'X1': 'gyro_x',
    'X2': 'gyro_y',
    'X3': 'gyro_z'
    })
  sf = sf.add_columns(gyro_sf)

  # calc labels
  exp_labels = labels[labels['exp_id'] == exp_id][['activity_id', 'start', 'end']].to_numpy()
  sf = sf.add_row_number()
  sf['activity_id'] = sf['id'].apply(
    lambda x: find_label_for_containing_interval(exp_labels, x)
  )
  sf = sf.remove_columns(['id', 'exp_id'])

  data = data.append(sf)

target_map = {
  1.: 'walking',
  2.: 'climbing_upstairs',
  3.: 'climbing_downstairs',
  4.: 'sitting',
  5.: 'standing',
  6.: 'laying'
}

# use the same labels used in the experiment
data = data.filter_by(target_map.keys(), 'activity_id')
data['activity'] = data['activity_id'].apply(
  lambda x: target_map[x]
)
data = data.remove_column('activity_id')

data.save('hapt_data.sframe')

生成的sframe文件:

sframe文件

具体每个文件是什么意思,还没搞懂...

根据生成的sframe训练core ml所需的mlmodel文件

import turicreate as tc

# load sessions from preprocessed data
data = tc.SFrame('hapt_data.sframe')

# train/test split by recording sessions
train, test = tc.activity_classifier.util.random_split_by_session(
  data,
  session_id='user_id',
  fraction=0.8
)

# create an activity classifier
model = tc.activity_classifier.create(
  train,
  session_id='user_id',
  target='activity',
  prediction_window=50
)

# evaluate the model and save the results into a dictionary
metrics = model.evaluate(test)
print metrics['accuracy']

# save the model for later use in turi create
model.save('mymodel.model')

# export for use in core ml
model.export_coreml('MyActivityClassifier.mlmodel')

train result

训练的准确率一般啦!!!

model的高级运用:

walking_3_sec = data[(data['activity'] == 'walking') & (data['user_id'] == 1)][50:200]

print model.predict(walking_3_sec, output_frequency='per_window')

print model.predict(walking_3_sec, output_frequency='per_row')

3.把model放到app中使用

建个工程,把mlmodel拉进去

project

然后按照文档说明把预测代码写到ViewController.swift里

import UIKit
import CoreML
import CoreMotion

class ViewController: UIViewController {
    let motionManager = CMMotionManager()
    
    struct ModelConstants {
        static let numOfFeatures = 6
        static let predictionWindowSize = 50
        static let sensorsUpdateInterval = 1.0 / 50.0
        static let hiddenInLength = 200
        static let hiddenCellInLength = 200
    }
    let activityClassificationModel = MyActivityClassifier()
    
    var currentIndexInPredictionWindow = 0
    let predictionWindowDataArray = try? MLMultiArray(
        shape: [1 , ModelConstants.predictionWindowSize , ModelConstants.numOfFeatures] as [NSNumber],
        dataType: MLMultiArrayDataType.double
    )
    var lastHiddenOutput = try? MLMultiArray(
        shape: [ModelConstants.hiddenInLength as NSNumber],
        dataType: MLMultiArrayDataType.double
    )
    var lastHiddenCellOutput = try? MLMultiArray(
        shape: [ModelConstants.hiddenCellInLength as NSNumber],
        dataType: MLMultiArrayDataType.double
    )
    
    override func viewDidLoad() {
        super.viewDidLoad()
        motionManager.accelerometerUpdateInterval = TimeInterval(ModelConstants.sensorsUpdateInterval)
        motionManager.gyroUpdateInterval = TimeInterval(ModelConstants.sensorsUpdateInterval);
        
        motionManager.startAccelerometerUpdates(to: .main) { accelerometerData, error in
            guard let accelerometerData = accelerometerData else { return }
            self.addAccelSampleToDataArray(accelSample: accelerometerData)
        }
    }
    
    func addAccelSampleToDataArray(accelSample: CMAccelerometerData) {
        // add the current accelermeter reading to the data array
        guard let dataArray = predictionWindowDataArray else {
            return
        }
        dataArray[[0, currentIndexInPredictionWindow, 0] as [NSNumber]] = accelSample.acceleration.x as NSNumber
        dataArray[[0, currentIndexInPredictionWindow, 1] as [NSNumber]] = accelSample.acceleration.y as NSNumber
        dataArray[[0, currentIndexInPredictionWindow, 2] as [NSNumber]] = accelSample.acceleration.z as NSNumber
        
        // update the index in the prediction window data array
        currentIndexInPredictionWindow += 1
        
        // if the data array if full, call the prediction method to get a new model prediction.
        // we assume here for simplicity that the gyro data was added to the data array as well.
        if (currentIndexInPredictionWindow == ModelConstants.predictionWindowSize) {
            let predictedActivity = performModelPrediction() ?? "N/A"
            
            // use the predicted activity here
            print(predictedActivity)
            
            // start a new prediction window
            currentIndexInPredictionWindow = 0
        }
    }
    
    func performModelPrediction() -> String? {
        guard let dataArray = predictionWindowDataArray else {
            return "error"
        }
        
        // perform model prediction
        let modelPrediction = try? activityClassificationModel.prediction(
            features: dataArray,
            hiddenIn: lastHiddenOutput,
            cellIn: lastHiddenCellOutput
        )
        
        // update the state vectors
        lastHiddenOutput = modelPrediction?.hiddenOut
        lastHiddenCellOutput = modelPrediction?.cellOut
        
        // return the predicted activity - the activity with the highest probability
        return modelPrediction?.activity
    }
}

代码这么简单,就不另外放demo地址了,把代码贴上去就可以直接跑了.

于是就可以打印出预测的接下来动作了

result

还是很准的!