Narrow Results

The smart wearable devices that can track the fitness activities are getting famous these days due to their easy-to-use features. The fitness trackers can work for an individual in a promising manner, provided that the user is well familiar with the device and is committed with the timelines. Several reports have provided evidence that these smart wearable devices have not showed promising results and in most of the cases, people have stopped using them, few weeks after the purchase. There are several reasons linked with this response. During this research, we have worked on the correlations of weight loss via smart device with the age, gender, body mass index (BMI) and ideal body weight (IBW), with the aid of gradient boosted decision trees (XGBoost) and support vector machine (SVM) learning tools. XGBoost and SVM are capable of dealing with complex datasets, with higher frequencies, and for data emerging from multiple sources. These machine learning tools use kernel functions for the clustering and other classification measures, and are thus better as compared to the logistic methods. Next, the time series forecasting tools are discussed with the Bayesian hyperparametric optimization. The time series of the weight loss monitoring of each individual, depicted in this manner, provided complex fractal patterns, with reduction in amplitude, with the passage of time.

To support people trying to lose weight and stay healthy, more and more fitness apps have sprung up including the ability to track both calories intake and expenditure. Users of such apps are part of a wider “quantified self” movement and many opt-in to publicly share their logged data. In this paper, we use public food diaries of more than 4,000 long-term active MyFitnessPal users to study the characteristics of a (un-)successful diet. Concretely, we train a machine learning model to predict repeatedly being over or under self-set daily calories goals and then look at which features contribute to the model's prediction. Our findings include both expected results, such as the token “mcdonalds” or the category “dessert” being indicative for being over the calories goal, but also less obvious ones such as the difference between pork and poultry concerning dieting success, or the use of the “quick added calories” functionality being indicative of over-shooting calorie-wise. This study also hints at the feasibility of using such data for more in-depth data mining, e.g., looking at the interaction between consumed foods such as mixing protein- and carbohydrate-rich foods. To the best of our knowledge, this is the first systematic study of public food diaries.