A Comparison Study of Classifier Algorithms for Mobile-phone's Accelerometer Based Activity Recognition

On the use of ensemble of classifiers for accelerometer-based activity recognition

“…Many machine learning approaches such as Hidden Markov Models (HMMs) [2], Decision Trees (DT) [3], Support Vector Machines (SVM) [4], Conditional Random Fields (CRFs) [5], knearest Neighbor (KNN) Ayu et al [6] were successfully used in AR studies. There are many sensors which can be used for AR problem, and some of them which were used previously are given as follows: accelerometers, gyroscopes, magnetometer, GPS, RFID, light sensor, Inertial measurement units, skin temperature, ECG, EEG, and camera [1].…”

Improving Accelerometer-Based Activity Recognition by Using Ensemble of Classifiers

“…Luštrek, et al proposed using sensor data from smartphone to recognize activity for diabetes patients. Nine algorithms have been used in Weka, the classification accuracy was 88% [10]. …”

ClusterNN

“…The paper identifies many systems that are based on using smartphone sensors for activity recognition. Also, a comparative study of different classifier algorithms from Weka [5] machine learning tool was performed in [2] using data obtained from smartphone accelerometer. The data collected with phone placed in the shirt pocket was used to compare accuracies of IBK, Naive Bayes, Rotation Forest, VFI, DTNB and LMT algorithms while the data collected when the phone was placed in the palm position was used to compare accuracies of SMO, NNge, ClassificationViaRegression, FT, VFI, IBK and Naive Bayes algorithms.…”