A New Learning Algorithm Based on Lever Principle

We propose to train trading systems and portfolios by optimizing objective functions that directly measure trading and investment performance. Rather than basing a trading system on forecasts or training via a supervised learning algorithm using labelled trading data, we train our systems using recurrent reinforcement learning (RRL) algorithms. The performance functions that we consider for reinforcement learning are profit or wealth, economic utility, the Sharpe ratio and our proposed differential Sharpe ratio. The trading and portfolio management systems require prior decisions as input in order to properly take into account the effects of transactions costs, market impact, and taxes. This temporal dependence on system state requires the use of reinforcement versions of standard recurrent learning algorithms. We present empirical results in controlled experiments that demonstrate the efficacy of some of our methods for optimizing trading systems and portfolios. For a long/short trader, we find that maximizing the differential Sharpe ratio yields more consistent results than maximizing profits, and that both methods outperform a trading system based on forecasts that minimize MSE. We find that portfolio traders trained to maximize the differential Sharpe ratio achieve better risk-adjusted returns than those trained to maximize profit. Finally, we provide simulation results for an S&P 500/TBill asset allocation system that demonstrate the presence of out-of-sample predictability in the monthly S&P 500 stock index for the 25 year period 1970 through 1994. Copyright © 1998 John Wiley & Sons, Ltd.

This paper focuses on multi-modal Information Perception (IP) for Soft Robotic Hands (SRHs) using Machine Learning (ML) algorithms. A flexible Optical Fiber-based Curvature Sensor (OFCS) is fabricated, consisting of a Light-Emitting Diode (LED), photosensitive detector, and optical fiber. Bending the roughened optical fiber generates lower light intensity, which reflecting the curvature of the soft finger. Together with the curvature and pressure information, multi-modal IP is performed to improve the recognition accuracy. Recognitions of gesture, object shape, size, and weight are implemented with multiple ML approaches, including the Supervised Learning Algorithms (SLAs) of K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Logistic Regression (LR), and the unSupervised Learning Algorithm (un-SLA) of K-Means Clustering (KMC). Moreover, Optical Sensor Information (OSI), Pressure Sensor Information (PSI), and Double-Sensor Information (DSI) are adopted to compare the recognition accuracies. The experiment results demonstrate that the proposed sensors and recognition approaches are feasible and effective. The recognition accuracies obtained using the above ML algorithms and three modes of sensor information are higher than 85 percent for almost all combinations. Moreover, DSI is more accurate when compared to single modal sensor information and the KNN algorithm with a DSI outperforms the other combinations in recognition accuracy.

Autonomous underwater vehicles (AUVs) have great advantages for activities in deep sea, and expected as the attractive tool. However, AUVs have various problems which should be solved. In this paper, the Self-Organizing Map (SOM) is applied as the clustering method for the navigation system. The SOM is known as one of the effective methods to extract the principle feature from many parameters and decrease the dimension of parameters. Through the competitive learning algorithms, the obtained is tuned to express specific features of the input signals. We have been investigating the possibility of navigation system based on SOM through simulations are experiments with an AUV called Twin-Burger. The learning algorithm of usual SOM is unsupervised learning. However, supervised learning algorithms should be introduced because the relationship between distances information and desirable behavior of the robot, that is, the relationship from inputs to outputs should be acquired and learned. In this paper, a supervised learning algorithm is introduced into SOM and a method to adapt the local to its environment by learning and evaluating the trajectory of robot is proposed. In the proposed method, the map is made static and digital value as teaching data. In order to include more information of environment in the initial map, the trajectories of robot are evaluated, and the evaluation is utilized in the learning process. This method enables the to have both the effect of dynamics of robot and environmental information. The efficiency of the method is investigated through the simulations and experiments.

The learning method is critical for obtaining good generalisation in neural networks with limited training data. The Standard BackPropagation (SBP) training algorithm suffers from several problems such as sensitivity to the initial conditions and very slow convergence. The aim of this work is to use Genetic Programming (GP) to discover new supervised learning algorithms which can overcome some of these problems. In previous research a new learning algorithms for the output layer has been discovered using GP. By comparing this with SBP on different problems better performance was demonstrated. This paper shows that GP can also discover better learning algorithms for the hidden layers to be used in conjunction with the algorithm previously discovered. Comparing these with SBP on different problems we show they p rovide better performances. This study indicates that there exist many supervised learning algorithms better than SBP and that GP can be used to discover them.

An evaluation and choice of learning algorithms is a current research area in data mining, artificial intelligence and pattern recognition, etc. Supervised learning is one of the tasks most frequently used in data mining. There are several learning algorithms available in machine learning field and new algorithms are being added in machine learning literature. There is a need for selecting the best suitable learning algorithm for a given data. With the information explosion of different learning algorithms and the changing data scenarios, there is a need of smart learning system. The paper shows one approach where past experiences learned are used to suggest the best suitable learner using 3 meta-features namely simple, statistical and information theoretic features. The system tests 38 UCI benchmark datasets from various domains using nine classifiers from various categories. It is observed that for 29 datasets, i.e., 76 % of datasets, both the predicted and actual accuracies directly match. The proposed approach is found to be correct for algorithm selection of these datasets. New proposed equation of finding classifier accuracy based on meta-features is determined and validated. The study compares various supervised learning algorithms by performing tenfold cross-validation paired t test. The work helps in a critical step in data mining for selecting the suitable data mining algorithm.

BACKGROUNDTraditional methods of developing predictive models in inflammatory bowel diseases (IBD) rely on using statistical regression approaches to deriving clinical scores such as the Crohn's disease (CD) activity index. However, traditional approaches are unable to take advantage of more complex data structures such as repeated measurements. Deep learning methods have the potential ability to automatically find and learn complex, hidden relationships between predictive markers and outcomes, but their application to clinical prediction in CD and IBD has not been explored previously.AIMTo determine and compare the utility of deep learning with conventional algorithms in predicting response to anti-tumor necrosis factor (anti-TNF) therapy in CD.METHODSThis was a retrospective single-center cohort study of all CD patients who commenced anti-TNF therapy (either adalimumab or infliximab) from January 1, 2010 to December 31, 2015. Remission was defined as a C-reactive protein (CRP) < 5 mg/L at 12 mo after anti-TNF commencement. Three supervised learning algorithms were compared: (1) A conventional statistical learning algorithm using multivariable logistic regression on baseline data only; (2) A deep learning algorithm using a feed-forward artificial neural network on baseline data only; and (3) A deep learning algorithm using a recurrent neural network on repeated data. Predictive performance was assessed using area under the receiver operator characteristic curve (AUC) after 10× repeated 5-fold cross-validation.RESULTSA total of 146 patients were included (median age 36 years, 48% male). Concomitant therapy at anti-TNF commencement included thiopurines (68%), methotrexate (18%), corticosteroids (44%) and aminosalicylates (33%). After 12 mo, 64% had CRP < 5 mg/L. The conventional learning algorithm selected the following baseline variables for the predictive model: Complex disease behavior, albumin, monocytes, lymphocytes, mean corpuscular hemoglobin concentration and gamma-glutamyl transferase, and had a cross-validated AUC of 0.659, 95% confidence interval (CI): 0.562-0.756. A feed-forward artificial neural network using only baseline data demonstrated an AUC of 0.710 (95%CI: 0.622-0.799; P = 0.25 vs conventional). A recurrent neural network using repeated biomarker measurements demonstrated significantly higher AUC compared to the conventional algorithm (0.754, 95%CI: 0.674-0.834; P = 0.036).CONCLUSIONDeep learning methods are feasible and have the potential for stronger predictive performance compared to conventional model building methods when applied to predicting remission after anti-TNF therapy in CD.

Abstract: This paper describes empirical research into the effectiveness of using concepts and principles of physics in the teaching of geometrical proofs. A unit on using the lever principle and the principle of the unique centre of gravity of a triangle as premises in mathematical proof was taught in a Grade 12 mathematics class. The students were then asked to prove twice that the three medians of a triangle meet at one point, first using traditional geometrical arguments and then using arguments from physics. The students then responded to a structured questionnaire and a follow‐up interview that asked them to compare the two methods of proving. Most of the students found the proof using arguments from physics clearer and easier to remember than the proof using arguments from geometry. The study demonstrated, tentatively, that proofs from physics could be a viable addition to the mathematics curriculum.

Nowadays, the renewable energy sector is an area of interest for every state. Global regulations and policies encourage the development of these technologies, given the current political context, but also environmental issues. Romania, due to its geographical position and climate, is considered a country with high potential regarding the implementation of alternative sources of renewable energy. This research presents the importance of solar energy and provides a statistical analysis on the sectors influencing the implementation of green energy. At the same time, those counties that are eligible are identified and different scenarios are created for the ineligible counties that lead to their eligibility. The research develops 3 main objectives. To begin with, it is desired to be created an overview of the indicators included in the analysis, in order to develop a detailed statistical analysis of the situation of each county of Romania. Following this extracted information, the second objective is outlined, which is to create an indicator that groups counties into counties eligible for solar energy and counties ineligible for solar energy using the K-Means Cluster-unsupervised learning algorithm. Finally, using the supervised learning algorithm - Logistic Regression, predictions will be made with the help of which those sectors of activity that can be improved in order to implement green energy will be identified.

The research presented in this paper aims to automatically detect the singing voice in traditional Tunisian music, taking into account the main characteristics of the sound of the voice in this particular music style. This means creating the possibility to automatically identify instrumental and singing sounds. Therefore different methods for the automatic classification of sounds using supervised learning algorithms were compared and evaluated. The research is divided into four successive stages. First, the extraction of features vectors from the audio tracks (through calculation of the parameters of sound perception) followed by the selection and transformation process of relevant features for singing/instrumental discrimination. Then, using learning algorithms, the instrumental and vocal classes were modeled from a manually annotated database. Finally, the evaluation of the decision-making process (indexing) was applied on the test part of the database. The musical databases used for this study consists of extracts from the national sound archives of Centre of Mediterranean and Arabic Music (CMAM) and recordings made especially for this research. The possibility to index audio data (classify/segment) into vocal and instrumental recognition allows for the retrieval of content-based information of musical databases.

Neuro-fuzzy [Jang,1997][Abraham,2005] are hybrid systems that combine the learning capacity of neural nets [Haykin,1999] with the linguistic interpretation of fuzzy inference systems [Ross,2004]. These systems have been evaluated quite intensively in machine learning tasks. This is mainly due to a number of factors: the applicability of learning algorithms developed for neural nets; the possibility of promoting implicit and explicit knowledge integration; and the possibility of extracting knowledge in the form of fuzzy rules. Most of the well known neuro-fuzzy systems, however, present limitations regarding the number of inputs allowed or the limited (or nonexistent) form to create their own structure and rules [Nauck,1997][Nauck,19 98][Vuorimaa,1994][Zhang,1995]. This paper describes a new class of neuro-fuzzy models, called Hierarchical Neuro-Fuzzy BSP Systems (HNFB). These models employ the BSP partitioning (Binary Space Partitioning) of the input space [Chrysanthou,1996] and have been developed to bypass traditional drawbacks of neuro-fuzzy systems. This paper introduces the HNFB models based on supervised learning algorithm. These models were evaluated in many benchmark applications related to classification and time-series forecasting. A second paper, entitled Hierarchical Neuro-Fuzzy Systems Part II, focuses on hierarchical neuro-fuzzy models based on reinforcement learning algorithms.

This work presents a comprehensive review of the developments in using Machine Learning (ML)-based algorithms for the modeling and design optimization of switched reluctance motors (SRMs). We reviewed Machine Learning-based numerical and analytical approaches used in modeling SRMs. We showed the difference between the supervised, unsupervised and reinforcement learning algorithms. More focus is placed on supervised learning algorithms as they are the most used algorithms in this area. The supervised learning algorithms studied in this work include the feedforward neural networks, recurrent neural networks, support vector machines, extreme learning machines, and Bayesian networks. This work also discusses several essential aspects of the considered machine learning algorithms, such as core concept, structure, and computational time. It also surveys sample data acquisition methods and data size. Finally, comparisons between the different considered ML-based algorithms are conducted in terms of electric motor type, dataset inputs and outputs, and algorithm&#x2019;s structure and accuracy to provide a summary overview of the ML-based algorithms for SRMs modeling and design.

Predicting the petrophysical properties of rock samples using micro-CT images has gained significant attention recently. However, an accurate and an efficient numerical tool is still lacking. After investigating three numerical techniques, (i) pore network modeling (PNM), (ii) the finite volume method (FVM), and (iii) the lattice Boltzmann method (LBM), a workflow based on machine learning is established for fast and accurate prediction of permeability directly from 3D micro-CT images. We use more than 1100 samples scanned at high resolution and extract the relevant features from these samples for use in a supervised learning algorithm. The approach takes advantage of the efficient computation provided by PNM and the accuracy of the LBM to quickly and accurately estimate rock permeability. The relevant features derived from PNM and image analysis are fed into a supervised machine learning model and a deep neural network to compute the permeability in an end-to-end regression scheme. Within a supervised learning framework, machine and deep learning algorithms based on linear regression, gradient boosting, and physics-informed convolutional neural networks (CNNs) are applied to predict the petrophysical properties of porous rock from 3D micro-CT images. We have performed the sensitivity analysis on the feature importance, hyperparameters, and different learning algorithms to make a prediction. Values of R2 scores up to 88% and 91% are achieved using machine learning regression models and the deep learning approach, respectively. Remarkably, a significant gain in computation time—approximately 3 orders of magnitude—is achieved by applied machine learning compared with the LBM. Finally, the study highlights the critical role played by feature engineering in predicting petrophysical properties using deep learning.

There has been research around the idea of representing words in text as vectors and many models proposed that vary in performance as well as applications. Text processing is used for content recommendation, sentiment analysis, plagiarism detection, content creation, language translation, etc. to name a few. Specifically, we want to look at the problem of topic detection in text content of articles/blogs/summaries. With the humungous amount of text content published each and every minute on the internet, it is imperative that we have very good algorithms and approaches to analyze all the content and be able to classify most of it with high confidence for further use. The project aims to work with unsupervised and supervised machine learning algorithms in an effort to tackle the topic detection problem. The project will target various unsupervised learning algorithms like Word2vec, doc2vec and LDA for corpus and language dictionary learning to have a trained model which understand the semantic of texts. The objective of the project is to combine this unsupervised learning with supervised learning algorithms like Support Vector Machine and deep learning methods to analyze and hopefully better the performance in terms of accuracy of topic detection. The project also aims at performing user interest-based modelling, which is orthogonal to topics modeling. The idea is to make sure the model is free of predefined categories. The project results show that hybrid models are comfortably accurate when classifying text in a particular topic category. The project also concludes that user interest modelling can also be accurately achieved along with topic detection. The project successfully determines these results without any meta information about the input text and purely based on the corpus of the input text. This makes the project framework really robust as it has no dependency on source of text, length of text or any other meta information about the text content.

The insufficient loading of vehicles, which leads to a low logistics load factor is a common problem in the logistics industry. This study addresses this issue by utilizing actual shipment data from an automotive company. An effective method has been proposed to improve the company’s logistics efficiency through a scenario-based approach. Two real- world scenarios were developed to enhance vehicle loading performance. Machine learning algorithms were employed to evaluate the shipment performance of these scenarios. For the study, a dataset was generated from the company’s ERP system and real-world shipment data. Since this is a real-world problem, the dataset consisted of unlabeled data. To solve this problem, both supervised and unsupervised learning algorithms were applied. First, unsupervised clustering algorithms were used to group the shipment performance based on similarities. Then, supervised learning algorithms were utilized to classify the data within each group. The average cost was used to evaluate the clusters obtained through the unsupervised methods, while classification performance was measured using supervised machine learning techniques. The scenario-based approach has significantly improved the performance of the shipments as it shows the changes in load factor more clearly. In the actual case, only %25.7 of shipments were high performing, while this percentage gradually increased to %98.4 in the scenarios. The results show that optimizing the load factor makes the transports more efficient and balanced.

BackgroundAsthma is a variable long-term condition. Currently, there is no cure for asthma and the focus is, therefore, on long-term management. Mobile health (mHealth) is promising for chronic disease management but to be able to realize its potential, it needs to go beyond simply monitoring. mHealth therefore needs to leverage machine learning to provide tailored feedback with personalized algorithms. There is a need to understand the extent of machine learning that has been leveraged in the context of mHealth for asthma management. This review aims to fill this gap.MethodsWe searched PubMed for peer-reviewed studies that applied machine learning to data derived from mHealth for asthma management in the last five years. We selected studies that included some human data other than routinely collected in primary care and used at least one machine learning algorithm.ResultsOut of 90 studies, we identified 22 relevant studies that were then further reviewed. Broadly, existing research efforts can be categorized into three types: 1) technology development, 2) attack prediction, 3) patient clustering. Using data from a variety of devices (smartphones, smartwatches, peak flow meters, electronic noses, smart inhalers, and pulse oximeters), most applications used supervised learning algorithms (logistic regression, decision trees, and related algorithms) while a few used unsupervised learning algorithms. The vast majority used traditional machine learning techniques, but a few studies investigated the use of deep learning algorithms.DiscussionIn the past five years, many studies have successfully applied machine learning to asthma mHealth data. However, most have been developed on small datasets with internal validation at best. Small sample sizes and lack of external validation limit the generalizability of these studies. Future research should collect data that are more representative of the wider asthma population and focus on validating the derived algorithms and technologies in a real-world setting.