Search results for: operator method

Authors: Payam Mousavi, Andrew L. Stewart, Huiwen You, Aryeh F. G. Fayerman

Abstract:

We present a decision-support tool to assist an operator in the detection and tracking of a suspect vehicle traveling to an unknown target destination. Multiple data sources, such as traffic cameras, traffic information, weather, etc., are integrated and processed in real-time to infer a suspect’s intended destination chosen from a list of pre-determined high-value targets. Previously, we presented our work in the detection and tracking of vehicles using traffic and airborne cameras. Here, we focus on the fusion and processing of that information to predict a suspect’s behavior. The network of cameras is represented by a directional graph, where the edges correspond to direct road connections between the nodes and the edge weights are proportional to the average time it takes to travel from one node to another. For our experiments, we construct our graph based on the greater Los Angeles subset of the Caltrans’s “Performance Measurement System” (PeMS) dataset. We propose a Bayesian approach where a posterior probability for each target is continuously updated based on detections of the suspect in the live video feeds. Additionally, we introduce the concept of ‘soft interventions’, inspired by the field of Causal Inference. Soft interventions are herein defined as interventions that do not immediately interfere with the suspect’s movements; rather, a soft intervention may induce the suspect into making a new decision, ultimately making their intent more transparent. For example, a soft intervention could be temporarily closing a road a few blocks from the suspect’s current location, which may require the suspect to change their current course. The objective of these interventions is to gain the maximum amount of information about the suspect’s intent in the shortest possible time. Our system currently operates in a human-on-the-loop mode where at each step, a set of recommendations are presented to the operator to aid in decision-making. In principle, the system could operate autonomously, only prompting the operator for critical decisions, allowing the system to significantly scale up to larger areas and multiple suspects. Once the intended target is identified with sufficient confidence, the vehicle is reported to the authorities to take further action. Other recommendations include a selection of road closures, i.e., soft interventions, or to continue monitoring. We evaluate the performance of the proposed system using simulated scenarios where the suspect, starting at random locations, takes a noisy shortest path to their intended target. In all scenarios, the suspect’s intended target is unknown to our system. The decision thresholds are selected to maximize the chances of determining the suspect’s intended target in the minimum amount of time and with the smallest number of interventions. We conclude by discussing the limitations of our current approach to motivate a machine learning approach, based on reinforcement learning in order to relax some of the current limiting assumptions.

Keywords: autonomous surveillance, Bayesian reasoning, decision support, interventions, patterns of life, predictive analytics, predictive insights

Procedia PDF Downloads 93

Authors: Salvatore Triolo

Abstract:

We discuss the possibility of extending the well-known Gelfand-Naimark-Segal representation to modules over a C*algebra. We focus our attention on the case of Hilbert modules. We consider, in particular, the problem of the existence of a faithful representation. Non-commutative Lᵖ-spaces are shown to constitute examples of a class of CQ*-algebras. Finally, we have shown that any semisimple proper CQ*-algebra (X, A#), with A# a W*-algebra can be represented as a CQ*-algebra of measurable operators in Segal’s sense.

Keywords: Gelfand-Naimark-Segal representation, CQ*-algebras, faithful representation, non-commutative Lᵖ-spaces, operator in Hilbert spaces

Procedia PDF Downloads 221

Authors: Kholod Mohammad Abualnaja

Abstract:

A new algorithm is presented to find some new iterative methods for solving nonlinear equations F(x)=0 by using the variational iteration method. The efficiency of the considered method is illustrated by example. The results show that the proposed iteration technique, without linearization or small perturbation, is very effective and convenient.

Keywords: variational iteration method, nonlinear equations, Lagrange multiplier, algorithms

Procedia PDF Downloads 513

Authors: M. S. Baazzim, M. S. Al-Saud, M. A. El-Kady

Abstract:

In this paper, steady-state ampacity (current carrying capacity) evaluation of underground power cable system by using analytical and numerical methods for different conditions (depth of cable, spacing between phases, soil thermal resistivity, ambient temperature, wind speed), for two system voltage level were used 132 and 380 kV. The analytical method or traditional method that was used is based on the thermal analysis method developed by Neher-McGrath and further enhanced by International Electrotechnical Commission (IEC) and published in standard IEC 60287. The numerical method that was used is finite element method and it was recourse commercial software based on finite element method.

Keywords: cable ampacity, finite element method, underground cable, thermal rating

Procedia PDF Downloads 343

Authors: P. K. Singhal, R. Naresh, V. Sharma

Abstract:

This paper presents the three optimization models, namely New Binary Artificial Bee Colony (NBABC) algorithm, NBABC with Local Search (NBABC-LS), and NBABC with Genetic Crossover (NBABC-GC) for solving the Wind-Thermal Unit Commitment (WTUC) problem. The uncertain nature of the wind power is incorporated using the Weibull probability density function, which is used to calculate the overestimation and underestimation costs associated with the wind power fluctuation. The NBABC algorithm utilizes a mechanism based on the dissimilarity measure between binary strings for generating the binary solutions in WTUC problem. In NBABC algorithm, an intelligent scout bee phase is proposed that replaces the abandoned solution with the global best solution. The local search operator exploits the neighboring region of the current solutions, whereas the integration of genetic crossover with the NBABC algorithm increases the diversity in the search space and thus avoids the problem of local trappings encountered with the NBABC algorithm. These models are then used to decide the units on/off status, whereas the lambda iteration method is used to dispatch the hourly load demand among the committed units. The effectiveness of the proposed models is validated on an IEEE 10-unit thermal system combined with a wind farm over the planning period of 24 hours.

Keywords: artificial bee colony algorithm, economic dispatch, unit commitment, wind power

Procedia PDF Downloads 352

Authors: M. S. H. Chowdhury, Ishak Hashim

Abstract:

In this paper, linear and non-linear stiff systems of ordinary differential equations are solved by the classical Adomian decomposition method (ADM) and the multi-stage Adomian decomposition method (MADM). The MADM is a technique adapted from the standard Adomian decomposition method (ADM) where standard ADM is converted into a hybrid numeric-analytic method called the multistage ADM (MADM). The MADM is tested for several examples. Comparisons with an explicit Runge-Kutta-type method (RK) and the classical ADM demonstrate the limitations of ADM and promising capability of the MADM for solving stiff initial value problems (IVPs).

Keywords: stiff system of ODEs, Runge-Kutta Type Method, Adomian decomposition method, Multistage ADM

Procedia PDF Downloads 403

Authors: L. Fridrichova, R. Knížek, V. Bajzík

Abstract:

Drape is one of the important visual characteristics of the fabric. This paper is introducing an innovative method of measurement and evaluation of the drape shape of the fabric. The measuring principle is based on the possibility of multiple vertical strain of the fabric. This method more accurately simulates the real behavior of the fabric in the process of draping. The method is fully automated, so the sample can be measured by using any number of cycles in any time horizon. Using the present method of measurement, we are able to describe the viscoelastic behavior of the fabric.

Keywords: drape, drape shape, automated drapemeter, fabric

Procedia PDF Downloads 624

Authors: N. Senu, I. A. Kasim, F. Ismail, N. Bachok

Abstract:

In this paper zero-dissipative explicit Runge-Kutta method is derived for solving second-order ordinary differential equations with periodical solutions. The phase-lag and dissipation properties for Runge-Kutta (RK) method are also discussed. The new method has algebraic order three with dissipation of order infinity. The numerical results for the new method are compared with existing method when solving the second-order differential equations with periodic solutions using constant step size.

Keywords: dissipation, oscillatory solutions, phase-lag, Runge-Kutta methods

Procedia PDF Downloads 377

Authors: Chui Ka Shing

Abstract:

This case study research attempts to examine the use of the Bar Model Method approach in learning and teaching mathematics in a primary school in Hong Kong. The objectives of the study are to find out to what extent (a) the Bar Model Method approach enhances the construction of students’ mathematics concepts, and (b) the school-based mathematics curriculum development with adopting the Bar Model Method approach. This case study illuminates the effectiveness of using the Bar Model Method to solve mathematics problems from Primary 1 to Primary 6. Some effective pedagogies and assessments were developed to strengthen the use of the Bar Model Method across year levels. Suggestions including school-based curriculum development for using Bar Model Method and further study were discussed.

Keywords: bar model method, curriculum development, mathematics education, problem solving

Procedia PDF Downloads 187

Authors: Hemad Zareiforoush, Saeed Minaei, Ahmad Banakar, Mohammad Reza Alizadeh

Abstract:

The quality inspection of rice (Oryza sativa L.) during its various processing stages is very important. In this research, an artificial intelligence-based model coupled with computer vision techniques was developed as a decision support system for qualitative grading of rice grains. For conducting the experiments, first, 25 samples of rice grains with different levels of percentage of broken kernels (PBK) and degree of milling (DOM) were prepared and their qualitative grade was assessed by experienced experts. Then, the quality parameters of the same samples examined by experts were determined using a machine vision system. A grading model was developed based on fuzzy logic theory in MATLAB software for making a relationship between the qualitative characteristics of the product and its quality. Totally, 25 rules were used for qualitative grading based on AND operator and Mamdani inference system. The fuzzy inference system was consisted of two input linguistic variables namely, DOM and PBK, which were obtained by the machine vision system, and one output variable (quality of the product). The model output was finally defuzzified using Center of Maximum (COM) method. In order to evaluate the developed model, the output of the fuzzy system was compared with experts’ assessments. It was revealed that the developed model can estimate the qualitative grade of the product with an accuracy of 95.74%.

Keywords: machine vision, fuzzy logic, rice, quality

Procedia PDF Downloads 381

Authors: Yang Zhong, Heng Liu

Abstract:

The decoupling method and the modified Naiver method are combined for accurate bending analysis of rectangular thick plates with all edges clamped supported. The basic governing equations for Mindlin plates are first decoupled into independent partial differential equations which can be solved separately. Using modified Navier method, the analytic solution of rectangular thick plate with all edges clamped supported is then derived. The solution method used in this paper leave out the complicated derivation for calculating coefficients and obtain the solution to problems directly. Numerical comparisons show the correctness and accuracy of the results at last.

Keywords: Mindlin plates, decoupling method, modified Navier method, bending rectangular plates

Procedia PDF Downloads 561

Authors: Azizakhanim Maharramli

Abstract:

The gradual transition from entrenched traditional technology and organization of construction production to innovative additive construction technology inevitably meets technological, technical, organizational, labour, and, finally, social difficulties. Therefore, the chosen nodal method will lead to the elimination of the above difficulties, combining some of the usual methods of construction and the myth in world practice that the labour force is subjected to a strong stream of reduction. The nodal method of additive technology will create favourable conditions for the optimal degree of distribution of labour across facilities due to the consistent performance of homogeneous work and the introduction of additive technology and traditional technology into construction production.

Keywords: parallel method, sequential method, stream method, combined method, nodal method

Procedia PDF Downloads 53

Authors: Mahacine Amrani

Abstract:

A simple and rapid method for the determination of alcohol in gasoline-alcohol mixtures using density measurements is described. The method can determine a minimum of 1% of alcohol by volume. The precision of the method is ± 3%.The method is more useful for field test in the quality assessment of alcohol blended fuels.

Keywords: gasoline-alcohol, mixture, alcohol determination, density, measurement, Morocco

Procedia PDF Downloads 279

Authors: Daniel Kostrzewa, Henryk Josiński

Abstract:

The expanded Invasive Weed Optimization algorithm (exIWO) is an optimization metaheuristic modelled on the original IWO version created by the researchers from the University of Tehran. The authors of the present paper have extended the exIWO algorithm introducing a set of both deterministic and non-deterministic strategies of individuals’ selection. The goal of the project was to evaluate the exIWO by testing its usefulness for solving some test instances of the traveling salesman problem (TSP) taken from the TSPLIB collection which allows comparing the experimental results with optimal values.

Keywords: expanded invasive weed optimization algorithm (exIWO), traveling salesman problem (TSP), heuristic approach, inversion operator

Procedia PDF Downloads 806

Authors: Melusi Khumalo, Anastacia Dlamini

Abstract:

In this paper, we consider a numerical solution for nonlinear Fredholm integral equations of the second kind. We work with uniform mesh and use the Lagrange polynomials together with the Galerkin finite element method, where the weight function is chosen in such a way that it takes the form of the approximate solution but with arbitrary coefficients. We implement the finite element method to the nonlinear Fredholm integral equations of the second kind. We consider the error analysis of the method. Furthermore, we look at a specific example to illustrate the implementation of the finite element method.

Keywords: finite element method, Galerkin approach, Fredholm integral equations, nonlinear integral equations

Procedia PDF Downloads 341

Authors: Jiankang Wang, Hongyang Yu

Abstract:

This paper proposes a portable online 3D modeling method. The method first utilizes a depth camera to collect data and compresses the depth data using a frame-by-frame lossless data compression method. The color image is encoded using the H.264 encoding format. After the cloud obtains the color image and depth image, a 3D modeling method based on bundlefusion is used to complete the 3D modeling. The results of this study indicate that this method has the characteristics of portability, online, and high efficiency and has a wide range of application prospects.

Keywords: 3D reconstruction, bundlefusion, lossless compression, depth image

Procedia PDF Downloads 50

Authors: Haijie Li, Xuping Zhang

Abstract:

This paper presents a computationally efficient method for the modeling of robot manipulators with flexible links and joints. This approach combines the Discrete Time Transfer Matrix Method with the Finite Segment Method, in which the flexible links are discretized by a number of rigid segments connected by torsion springs; and the flexibility of joints are modeled by torsion springs. The proposed method avoids the global dynamics and has the advantage of modeling non-uniform manipulators. Experiments and simulations of a single-link flexible manipulator are conducted for verifying the proposed methodologies. The simulations of a three-link robot arm with links and joints flexibility are also performed.

Keywords: flexible manipulator, transfer matrix method, linearization, finite segment method

Procedia PDF Downloads 404

Authors: Lal Hussain, Wajid Aziz, Sajjad Ahmed Nadeem, Saeed Arif Shah, Abdul Majid

Abstract:

Brain electrical activity as reflected in Electroencephalography (EEG) have been analyzed and diagnosed using various techniques. Among them, complexity measure, nonlinearity, disorder, and unpredictability play vital role due to the nonlinear interconnection between functional and anatomical subsystem emerged in brain in healthy state and during various diseases. There are many social and economical issues of alcoholic abuse as memory weakness, decision making, impairments, and concentrations etc. Alcoholism not only defect the brains but also associated with emotional, behavior, and cognitive impairments damaging the white and gray brain matters. A recently developed signal analysis method i.e. Multiscale Permutation Entropy (MPE) is proposed to estimate the complexity of long-range temporal correlation time series EEG of Alcoholic and Control subjects acquired from University of California Machine Learning repository and results are compared with MSE. Using MPE, coarsed grained series is first generated and the PE is computed for each coarsed grained time series against the electrodes O1, O2, C3, C4, F2, F3, F4, F7, F8, Fp1, Fp2, P3, P4, T7, and T8. The results computed against each electrode using MPE gives higher significant values as compared to MSE as well as mean rank differences accordingly. Likewise, ROC and Area under the ROC also gives higher separation against each electrode using MPE in comparison to MSE.

Keywords: electroencephalogram (EEG), multiscale permutation entropy (MPE), multiscale sample entropy (MSE), permutation entropy (PE), mann whitney test (MMT), receiver operator curve (ROC), complexity measure

Procedia PDF Downloads 462

Authors: Ahmed Guerine, Ali El Hafidi, Bruno Martin, Philippe Leclaire

Abstract:

In this paper, we propose a method for the dynamic response of multi-storey structures with uncertain-but-bounded parameters. The effectiveness of the proposed method is demonstrated by a numerical example of three-storey structures. This equation is integrated numerically using Newmark’s method. The numerical results are obtained by the proposed method. The simulation accounting the interval analysis method results are compared with a probabilistic approach results. The interval analysis method provides a mean curve that is between an upper and lower bound obtained from the probabilistic approach.

Keywords: multi-storey structure, dynamic response, interval analysis method, random parameters

Procedia PDF Downloads 159

Authors: Somaya Adwan, Rasha Majed, Lamya'a Majed, Hamzah Arof

Abstract:

In order to produce images with whole body parts, X-ray of different portions of the body parts is assembled using image stitching methods. A new method for image stitching that exploits mutually feature based method and direct based method to identify and merge pairs of X-ray medical images is presented in this paper. The performance of the proposed method based on this hybrid approach is investigated in this paper. The ability of the proposed method to stitch and merge the overlapping pairs of images is demonstrated. Our proposed method display comparable if not superior performance to other feature based methods that are mentioned in the literature on the standard databases. These results are promising and demonstrate the potential of the proposed method for further development to tackle more advanced stitching problems.

Keywords: image stitching, direct based method, panoramic image, X-ray

Procedia PDF Downloads 514

Authors: Siamak Feizi, Gunvor Baardvik

Abstract:

Soil erosion has special consequences for landfills that are more serious than those found at conventional construction sites. Different potential heads between two sides of a landfill and the subsequent movement of water through pores within the soil body could trigger the soil erosion and construction instability. Such a condition was encountered in a landfill project in the southern part of Norway. To check the risk of internal erosion due to changes in the groundwater level (because of seasonal flooding in the river), a series of numerical simulations by means of Geo-Seep software was conducted. Output of this study provides a total picture of the landfill stability, possibilities of erosions, and necessary measures to prevent or reduce the risk for the landfill operator.

Keywords: erosion, seepage, landfill, stability

Procedia PDF Downloads 106

Authors: M. O. Olayiwola

Abstract:

Numerical solutions of the generalized Burger-Fisher are obtained using a Modified Variational Iteration Method (MVIM) with minimal computational efforts. The computed results with this technique have been compared with other results. The present method is seen to be a very reliable alternative method to some existing techniques for such nonlinear problems.

Keywords: burger-fisher, modified variational iteration method, lagrange multiplier, Taylor’s series, partial differential equation

Procedia PDF Downloads 402

Authors: Remzi Saltoglu, Nazmia Humaira, Gokhan Inalhan

Abstract:

During aircraft maintenance scheduling, operator calculates the budget of the maintenance. Usually, this calculation includes only the costs that are directly related to the maintenance process such as cost of labor, material, and equipment. In some cases, overhead cost is also included. However, in some of those, downtime cost is neglected claiming that grounding is a natural fact of maintenance; therefore, it is not considered as part of the analytical decision-making process. Based on the normalized data, we introduce downtime cost with its monetary value and add its seasonal character. We envision that the rest of the model, which works together with the downtime cost, could be checked with the real life cases, through the review of MRO cost and airline spending in the particular and scheduled maintenance events.

Keywords: aircraft maintenance, downtime, downtime cost, maintenance cost

Procedia PDF Downloads 326

Authors: Armanul Hoque Shaon, Md Baizid Mahmud, Askander Nobi, Md. Raju Ahmed, Md. Jiabul Hoque

Abstract:

In the Internet of Things (IoT), devices are linked to the internet through a wireless network, allowing them to collect and transmit data without the need for a human operator. Agriculture relies heavily on wireless sensors, which are a vital component of the Internet of Things (IoT). This kind of wireless sensor network monitors physical or environmental variables like temperatures, sound, vibration, pressure, or motion without relying on a central location or sink and collaboratively passes its data across the network to be analyzed. As the primary source of plant nutrients, the soil is critical to the agricultural industry's continued growth. We're excited about the prospect of developing an Internet of Things (IoT) solution. To arrange the network, the sink node collects groundwater levels and sends them to the Gateway, which centralizes the data and forwards it to the sensor nodes. The sink node gathers soil moisture data, transmits the mean to the Gateways, and then forwards it to the website for dissemination. The web server is in charge of storing and presenting the moisture in the soil data to the web application's users. Soil characteristics may be collected using a networked method that we developed to improve rice production. Paddy land is running out as the population of our nation grows. The success of this project will be dependent on the appropriate use of the existing land base.

Keywords: IoT based agriculture monitoring, intelligent irrigation, communicating network, rice production

Procedia PDF Downloads 125

Authors: Safeer Hussain Khan

Abstract:

In this paper, we introduce a two-step iterative algorithm to prove a strong convergence result for approximating common fixed points of three contractive-like operators. Our algorithm basically generalizes an existing algorithm..Our iterative algorithm also contains two famous iterative algorithms: Mann iterative algorithm and Ishikawa iterative algorithm. Thus our result generalizes the corresponding results proved for the above three iterative algorithms to a class of more general operators. At the end, we remark that nothing prevents us to extend our result to the case of the iterative algorithm with error terms.

Keywords: contractive-like operator, iterative algorithm, fixed point, strong convergence

Procedia PDF Downloads 516

Authors: Y. Sunecher, N. Mamode Khan, V. Jowaheer

Abstract:

This paper considers the modelling of a non-stationary bivariate integer-valued autoregressive moving average of order one (BINARMA(1,1)) with correlated Poisson innovations. The BINARMA(1,1) model is specified using the binomial thinning operator and by assuming that the cross-correlation between the two series is induced by the innovation terms only. Based on these assumptions, the non-stationary marginal and joint moments of the BINARMA(1,1) are derived iteratively by using some initial stationary moments. As regards to the estimation of parameters of the proposed model, the conditional maximum likelihood (CML) estimation method is derived based on thinning and convolution properties. The forecasting equations of the BINARMA(1,1) model are also derived. A simulation study is also proposed where BINARMA(1,1) count data are generated using a multivariate Poisson R code for the innovation terms. The performance of the BINARMA(1,1) model is then assessed through a simulation experiment and the mean estimates of the model parameters obtained are all efficient, based on their standard errors. The proposed model is then used to analyse a real-life accident data on the motorway in Mauritius, based on some covariates: policemen, daily patrol, speed cameras, traffic lights and roundabouts. The BINARMA(1,1) model is applied on the accident data and the CML estimates clearly indicate a significant impact of the covariates on the number of accidents on the motorway in Mauritius. The forecasting equations also provide reliable one-step ahead forecasts.

Keywords: non-stationary, BINARMA(1, 1) model, Poisson innovations, conditional maximum likelihood, CML

Procedia PDF Downloads 102

Authors: Adel A. Ghobbar, Ahmad Bakkar

Abstract:

Although special care is given to maintenance, aircraft systems fail, and these failures cause delays and cancellations. The occurrence of Delays and Cancellations affects operators and manufacturers negatively. To reduce technical delays and cancellations, one should be able to determine the important systems causing them. The goal of this research is to find a method to define the most expensive delays and cancellations systems for Airline operators. A predictive model was introduced to forecast the failure and their impact after carrying out research that identifies relevant information to tackle the problems faced while answering the questions of this paper. Data were obtained from the manufacturers’ services reliability team database. Subsequently, delays and cancellations evaluation methods were identified. No cost estimation methods were used due to their complexity. The model was developed, and it takes into account the frequency of delays and cancellations and uses weighting factors to give an indication of the severity of their duration. The weighting factors are based on customer experience. The data Analysis approach has shown that delays and cancellations events are not seasonal and do not follow any specific trends. The use of weighting factor does have an influence on the shortlist over short periods (Monthly) but not the analyzed period of three years. Landing gear and the navigation system are among the top 3 factors causing delays and cancellations for all three aircraft types. The results did confirm that the cooperation between certain operators and manufacture reduce the impact of delays and cancellations.

Keywords: reliability, availability, delays & cancellations, aircraft maintenance

Procedia PDF Downloads 103

Authors: Mohammad Ahmad, Dayalan Kasilingam

Abstract:

In this paper, a spectral domain implementation of the fast multipole method is presented. It is shown that the aggregation, translation, and disaggregation stages of the fast multipole method (FMM) can be performed using the spectral domain (SD) analysis. The spectral domain fast multipole method (SD-FMM) has the advantage of eliminating the near field/far field classification used in conventional FMM formulation. The study focuses on the application of SD-FMM to one-dimensional (1D) and two-dimensional (2D) electric field integral equation (EFIE). The case of perfectly conducting strip, circular and square cylinders are numerically analyzed and compared with the results from the standard method of moments (MoM).

Keywords: electric field integral equation, fast multipole method, method of moments, wave scattering, spectral domain

Procedia PDF Downloads 371

Authors: Alkaben Patel

Abstract:

The proposed RP-HPLC method is easy, rapid, economical, precise and accurate stability indicating RP-HPLC method for simultaneous estimation of Astorvastatin and Methylcobalamine in their combined dosage form has been developed.The separation was achieved by LC-20 AT C18(250mm*4.6mm*2.6mm)Colum and water (pH 3.5): methanol 70:30 as mobile phase, at a flow rate of 1ml/min. wavelength of this dosage form is 215nm.The drug is related to stress condition of hydrolysis, oxidation, photolysis and thermal degradation.

Keywords: RP- HPLC, atorvastatin, methylcobalamine, method, development, validation

Procedia PDF Downloads 303

Authors: U. Borasi, Amit Kr. Jain, Rakesh, Piyush Jain

Abstract:

Data collection, data preparation, model training, model evaluation, and deployment are all processes in a typical machine learning workflow. Training data needs to be gathered and organised. This often entails collecting a sizable dataset and cleaning it to remove or correct any inaccurate or missing information. Preparing the data for use in the machine learning model requires pre-processing it after it has been acquired. This often entails actions like scaling or normalising the data, handling outliers, selecting appropriate features, reducing dimensionality, etc. This pre-processed data is then used to train a model on some machine learning algorithm. After the model has been trained, it needs to be assessed by determining metrics like accuracy, precision, and recall, utilising a test dataset. Every time a new model is built, both data pre-processing and model training—two crucial processes in the Machine learning (ML) workflow—must be carried out. Thus, there are various Machine Learning algorithms that can be employed for every single approach to data pre-processing, generating a large set of combinations to choose from. Example: for every method to handle missing values (dropping records, replacing with mean, etc.), for every scaling technique, and for every combination of features selected, a different algorithm can be used. As a result, in order to get the optimum outcomes, these tasks are frequently repeated in different combinations. This paper suggests a simple architecture for organizing this largely produced “combination set of pre-processing steps and algorithms” into an automated workflow which simplifies the task of carrying out all possibilities.

Keywords: machine learning, automation, AUTOML, architecture, operator pool, configuration, scheduler

Procedia PDF Downloads 26