Gait Pattern Recognition in Patients with Hereditary Ataxia’s Using Supervised Learning Algorithms Based on Small Subsets Smartphone Sensor Data<b></b>

Abstract

The progressive impairment analysis in gait from neurological diseases patients such as Hereditary Ataxias (HA) has been carried out using gait data collected with movement sensors. This research is focused on finding the minimum amount required of gait features to recognize efficiently and less intrusive way, HA patients based on data collected with iPhone movement sensors placed on the ankles from 14 HA patients and 14 healthy people. A twofold proposal is made , first a local minimum prominent peak criterion to find out the starting point of each stride, to get 10-stride window about which 56 spatial-temporal features are derived; second a search strategy based on Hill Climbing algorithm to reduce the number of gait features and sensors. The main results were the findings that with two gait patterns a 96% of classification accuracy was achieved by using K-Nearest Neighbors (KNN) and Multi-Layer Perceptron (MLP) algorithms, but in addition, MLP only right ankle sensor patterns were required which also allows to reduce the intrusion.

Ataxic person prediction using feature optimized based on machine learning model

“…The data pre-processed with distributed stochastic neighbour embedding combine with RF algorithm achieves a higher accuracy of 98%. In [33] classify hereditary ataxia using gait features collected from motion sensors. They focused in establishing minimum gait feature required to perform hereditary ataxia classification using ankle-based motions sensors.…”

Machine Learning Algorithms for Classification Patients with Parkinson`s Disease and Hereditary Ataxias

“…Then, As accelerometers measure the rate of change of the velocity of an object (acceleration), the average acceleration of a motionless accelerometer should be equal to the acceleration of gravity; however, due to their high sensitivity, the Gait data captured by these devices is consistently fluctuating as the acceleration values change. Therefore, we employed zero normalization to eliminate the constant signal effects on each ax [23].…”

“…Research indicates that the iPhone's sensors are dependable and accurate enough to analyze and detect kinematic gait patterns [21], [22]. In addition, research related to gait evaluation and healthcare has shown that iPhone sensors can obtain quantified gait characteristics with adequate precision and consistency, notably in ankle position and in a manner that is easy, portable, and wearable [23]. This article should be viewed as expanding the conference paper published in [20].…”

Ataxia severity classification using enhanced feature selection and ranking optimization through machine learning model

“…From data collected on 43 subjects (20 healthy subjects, 23 ataxic subjects), resulting in 418 sequences of normal gait pattern, the maximum accuracy of 98% was attained using an RF classifier pre-processed by t-distributed-stochastic-neighbor segmentation to distinguish among healthy persons and individuals with ataxic-gait. Researchers have used motion sensor information to analyze how individuals with neurological disorders, like hereditary-ataxias (HA), walk over time [26]. By collecting data across 14 HA participants and 14 healthy individuals via iPhone motion sensors attached to their ankles, this study seeks to determine the minimum needed gait traits necessary for effective and less invasive HA patient recognition.…”

“…Next, in iterative manner considering configured 𝑆, the below steps are executed ∀ 𝑙 = 1 to 𝐿, where 𝐿 represent the grid size used during optimization process. Ataxia severity classification prediction model construction through (26);…”