Yakup Kara

Predicting direction of stock price index movement using artificial neural networks and support vector machines

Prediction of stock price index movement is regarded as a challenging task of financial time series prediction. An accurate prediction of stock price movement may yield profits for investors. Due to the complexity of stock market data, development of efficient models for predicting is very difficult. This study attempted to develop two efficient models and compared their performances in predicting the direction of movement in the daily Istanbul Stock Exchange (ISE) National 100 Index. The models are based on two classification techniques, artificial neural networks (ANN) and support vector machines (SVM). Ten technical indicators were selected as inputs of the proposed models. Two comprehensive parameter setting experiments for both models were performed to improve their prediction performances. Experimental results showed that average performance of ANN model (75.74%) was found significantly better than that of SVM model (71.52%).

Predicting bank financial failures using neural networks, support vector machines and multivariate statistical methods: A comparative analysis in the sample of savings deposit insurance fund (SDIF) transferred banks in Turkey

Bank failures threaten the economic system as a whole. Therefore, predicting bank financial failures is crucial to prevent and/or lessen the incoming negative effects on the economic system. This is originally a classification problem to categorize banks as healthy or non-healthy ones. This study aims to apply various neural network techniques, support vector machines and multivariate statistical methods to the bank failure prediction problem in a Turkish case, and to present a comprehensive computational comparison of the classification performances of the techniques tested. Twenty financial ratios with six feature groups including capital adequacy, asset quality, management quality, earnings, liquidity and sensitivity to market risk (CAMELS) are selected as predictor variables in the study. Four different data sets with different characteristics are developed using official financial data to improve the prediction performance. Each data set is also divided into training and validation sets. In the category of neural networks, four different architectures namely multi-layer perceptron, competitive learning, self-organizing map and learning vector quantization are employed. The multivariate statistical methods; multivariate discriminant analysis, k-means cluster analysis and logistic regression analysis are tested. Experimental results are evaluated with respect to the correct accuracy performance of techniques. Results show that multi-layer perceptron and learning vector quantization can be considered as the most successful models in predicting the financial failure of banks.

Balancing and sequencing mixed-model just-in-time U-lines with multiple objectives

This study deals with the mixed-model U-lines utilized in just-in-time (JIT) production systems. Successful implementations of mixed-model U-lines requires solutions to two important problems called line balancing and model sequencing. In terms of some balance-dependent performance measures the effectiveness of a mixed-model U-line can be increased by solving line balancing and model sequencing problems simultaneously. However, this may lead to inefficient values of sequence-dependent performance measures. Hence, increasing the effectiveness of a mixed-model U-line requires balancing and sequencing problems that be dealt with multiple objectives. Balancing and sequencing mixed-model U-lines with multiple objectives has not been considered in the literature to date. In this study, a multi-objective approach for balancing and sequencing mixed-model U-lines to simultaneously minimize the absolute deviations of workloads across workstations, part usage rate, and cost of setups is presented. To increase the performance of the proposed algorithm, a newly developed neighbourhood generation method is also employed. Since the performance measures considered in the study are conflicting with each other, the proposed algorithm suggests much flexibility and more realistic results to decision makers. Solution methodology is illustrated using an example and a two-stage comprehensive experimental study is conducted to determine the effective values of algorithm parameters and investigate the relationships between performance measures. Results show that the proposed approach is more realistic than the limited number of existing methodologies. The proposed approach is also extended to consider the stochastic completion times of tasks.

Binary fuzzy goal programming approach to single model straight and U-shaped assembly line balancing

Assembly line balancing generally requires a set of acceptable solutions to the several conflicting objectives. In this study, a binary fuzzy goal programming approach is applied to assembly line balancing. Models for balancing straight and U-shaped assembly lines with fuzzy goals (the number of workstations and cycle time goals) are proposed. The binary fuzzy goal programming models are solved using the methodology introduced by Chang [Chang, C.T., 2007. Binary fuzzy goal programming. European Journal of Operational Research 180 (1), 29-37]. An illustrative example is presented to demonstrate the validity of the proposed models and to compare the performance of straight and U-shaped line configurations.

A mixed integer linear programming formulation for optimal balancing of mixed-model U-lines

The mixed-model U-line balancing problem was first studied by Sparling and Miltenburg (Sparling, D. and Miltenburg, J., 1998. The mixed-model U-line balancing problem. International Journal of Production Research, 36(2), 485–501) but has not been mathematically formulated to date. This paper presents a mixed integer programming formulation for optimal balancing of mixed-model U-lines. The proposed approach minimises the number of workstations required on the line for a given model sequence. The proposed formulation is illustrated and tested on an example problem and compared with an existing approach. This paper also proposes a new heuristic solution procedure to handle large scale mixed-model U-line balancing problems. A comprehensive experimental analysis is also conducted to evaluate the performance of the proposed heuristic. The results show the validity and usefulness of the proposed integer formulation and effectiveness of the proposed heuristic procedure.

Line balancing and model sequencing to reduce work overload in mixed-model U-line production environments

Mixed-model U-lines (MMULs) are important elements of just-in-time production systems. For successful implementation of MMULs, a smoothed workload distribution among workstations is important. This requires that line balancing and model sequencing problems are solved simultaneously. This article presents a mixed, zero–one, nonlinear mathematical programming formulation for balancing and sequencing MMULs simultaneously with the objective of reducing work overload. Since the problem is NP-hard, an effective simulated annealing approach is also presented and its performance compared with existing approaches. The results show that the proposed simulated annealing algorithm outperforms existing approaches.

Balancing parallel assembly lines with precise and fuzzy goals

Balancing parallel assembly lines provides flexibility to minimise the total idle times of assembly lines and the total number of workstations required in production facilities. This paper proposes two goal programming approaches to balance parallel assembly lines with precise and fuzzy goals. Three conflicting goals, namely number of workstations, cycle time, and number of tasks assigned to a workstation are optimised in crisp and fuzzy environments. The proposed approaches are also illustrated using numerical examples and sensitivity analyses. The proposed approaches provide flexibility for decision makers to balance parallel assembly lines based on their decision environments and preferred priorities.

Balancing straight and U-shaped assembly lines with resource dependent task times

The basic assumption of assembly line balancing is that every tasks time is fixed. However, in practice, different processing alternatives may be available to process a task with different times. The problem in this case is to assign tasks and resources to work stations that minimise total cost and so-called resource dependent assembly line balancing (RDALB) in the literature. This study proposes a new integer programming formulation for RDALB. This formulation is then modified to develop a formulation for RDULB problem in U-shaped assembly lines. To the best knowledge of the authors, this study is the first RDULB study and the developed formulation is the first RDULB formulation. The proposed formulations are illustrated and validated using several examples. An experimental analysis is also conducted to examine the percent improvement in total cost when the line layout is switched into the U-shaped from the straight line shape. Experimental results show that an improvement in total operating cost is obtained when the straight line is switched into the U-shaped line configuration. Results also show that percentage improvement in total operating cost is significantly greater for problem instances having a large number of tasks and having greater values for the strength of the precedence relationships among tasks.

Parameter setting of the Fuzzy ART neural network to part-machine cell formation problem

The first step in cellular manufacturing system applications is the solution of the cell-formation problem. Many researchers have investigated this problem and a number of methods developed. In the present study, one of these methods, the Fuzzy ART neural network, was investigated. This method can solve the cell-formation problem using both binary and non-binary part-machine incidence matrices. The neural network model was coded in PASCAL and applied to 26 test problems to determine the efficient parameter combinations. Results show that the Fuzzy ART neural network can solve both binary and non-binary problems effectively. Results also show that’ parameter combinations for binary problems differ from parameter combinations for non-binary problems.

An integrated model to incorporate ergonomics and resource restrictions into assembly line balancing

This study incorporates ergonomics and resource restrictions into assembly line balancing (ALB). For this purpose, an integrated model is proposed. The proposed model is essentially a cost-oriented formulation for ALB under psychological strain, physical strain, worker skills, multiple workers, equipment, working postures and illumination level restrictions. The integrated nature of the proposed model is one of its distinctive features in ALB literature. In addition, the proposed model modifies the existing restrictions by means of real-life facts and introduces two new restrictions arisen from the industry that have not been addressed in the literature. The model minimises the overall cost associated with operating costs of workforce and resources utilised under the ergonomics and resource restrictions. The model is illustrated and validated using some examples.