Saeed VARASTEH YAZDI

The current pattern of food production and consumption in the world has put great pressure on the environment, so a shift towards sustainable diets seems more urgent than ever. Within this context, organizations such as hospitals offer valuable opportunities to promote sustainable eating practices. The aim of the present study is to design a menu that minimizes environmental impact and cost, while simultaneously maximizing nutrient-rich food (NRF). A two-phase linear optimization was conducted at Imam Reza Hospital, Mashhad, Iran, combining linear and goal programming to optimize the quantities of each food. In this regard, data on recipes, quantities, and prices were collected over a 461-day period in the hospital canteen. For each meal, NRF, water footprint, carbon footprint, and cost were calculated. Four different scenarios were developed, and the best scenario was used to design one-year meal plans that meet macronutrient targets for protein, fat, and carbohydrates, and maintain dietary variety. The optimized menu attained a 36% reduction in carbon footprint and a 42% reduction in water footprint, while enhancing NRF by 10% and reducing cost by 42% in comparison with the prevailing menu. The mean percentage change in fat and carbohydrates from the obtained menu was -7.5% and 6%, respectively, which ensured the mean percentage change in all macronutrient levels aligns with recommended guidelines. The results of this study showed that healthier menus can be designed with a smaller environmental footprint, while maintaining or even significantly reducing costs.

"Kernel methods are known to be effective to analyse complex objects by implicitly embedding them into some feature space. To interpret and analyse the obtained results, it is often required to restore in the input space the results obtained in the feature space, by using pre-image estimation methods. This work proposes a new closed-form pre-image estimation method for time series kernel analytics that consists of two steps. In the first step, a time warp function, driven by distance constraints in the feature space, is defined to embed time series in a metric space where analytics can be performed conveniently. In the second step, the time series pre-image estimation is cast as learning a linear (or a nonlinear) transformation that ensures a local isometry between the time series embedding space and the feature space. The proposed method is compared to the state of the art through three major tasks that require pre-image estimation: 1) time series averaging, 2) time series reconstruction and denoising and 3) time series representation learning. The extensive experiments conducted on 33 publicly-available datasets show the benefits of the pre-image estimation for time series kernel analytics."

"Accuracy of the k-nearest neighbour (kNN) classifier depends strongly on the ability of the used distance to induce k-nearest neighbours of the same class while keeping distant samples of different classes. For time series classification, kNN based on dynamic time warping (dtw) measure remains among the most popular and competitive approaches. However, by assuming time series uniformly distributed, standard dtw may show some limitations to classify complex time series. In this paper, we show how to enhance the potential of kNN under time warp measure by learning a locally weighted dynamic time warping. For that, first discriminative features are learned from the neighbourhoods, then used to weight time series elements to bring closer the k-nearest neighbours of the same class and move away the k-nearest neighbours of different classes. To evaluate the proposed method, a deep analysis and experimentation are conducted on 87 public datasets from different application domains, varying sizes and difficulty levels. The results obtained show significant improvement in the proposed weighted dtw for time series kNN classification."

This work proposes a time warp invariant sparse coding and dictionary learning framework for time series clustering, where both input samples and atoms define time series of different lengths that involve variable delays. For that, first an l0 sparse coding problem is formalised and a time warp invariant orthogonal matching pursuit based on a new cosine maximisation time warp operator is proposed. A dictionary learning under time warp is then formalised and a gradient descent solution is developed. Lastly, a time series clustering based on the time warp sparse coding and dictionary learning is presented. The proposed approach is evaluated and compared to major alternative methods on several public datasets, with an application to deezer music data stream clustering.

"Learning dictionary for sparse representing time series is an important issue to extract latent temporal features, reveal salient primitives and sparsely represent complex temporal data. Time series are challenging data, they are often of different durations, may be composed of local or global salient events, that may arise with varying delays at different time stamps. This paper addresses the sparse coding and dictionary learning for such challenging time series. For that, we propose a non linear time warp invariant kSVD (twi-ksvd) where both input samples and dictionary atoms may have different lengths while involving varying delays. For the sparse coding problem, we propose an efficient time warp invariant orthogonal matching pursuit based on a new cosine maximisation time warp operator and induced sibling atoms. For the dictionary learning, thanks to a rotation transformation between each atom and its sibling atoms, a singular value decomposition is used to jointly approximate the coefficients and update the dictionary. The proposed method is confronted to major shift invariant, convolved and kernel dictionary learning methods on several challenging character and digit handwritten trajectories. The experiments conducted show the potential of twi-ksvd to efficiently sparse represent time series and to extract latent discriminative primitives for time series classification."

"The definition of a metric between time series is inherent to several data analysis and mining tasks, including clustering, classification or forecasting. Time series data present naturally several modalities covering their amplitude, behavior or frequential spectrum, that may be expressed with varying delays and at multiple temporal scales—exhibited globally or locally. Combining several modalities at multiple temporal scales to learn a holistic metric is a key challenge for many real temporal data applications. This paper proposes a Multi-modal and Multi-scale Temporal Metric Learning (m2tml) approach for a robust time series nearest neighbors classification. The solution lies in embedding time series into a dissimilarity space where a pairwise svm is used to learn both linear and non linear combined metric. A sparse and interpretable variant of the solution shows the ability of the learned temporal metric to localize accurately discriminative modalities as well as their temporal scales. A wide range of 30 public and challenging datasets, encompassing images, traces and ecg data, are used to show the efficiency and the potential of m2tml for an effective time series nearest neighbors classification."