Purpose – The purpose of this paper is to propose data mining techniques to model the return on investment from various types of promotional spending to market a drug and then use the model to draw conclusions on how the pharmaceutical industry might go about allocating promotion expenditures in a more efficient manner, potentially reducing costs to the consumer. The main contributions of the paper are two-fold. First, it demonstrates how to undertake a promotion mix optimization process in the pharmaceutical context and carry it through from the beginning to the end. Second, the paper proposes using directed acyclic graphs (DAGs) to help unravel the direct and indirect effects of various promotional media on sales volume. Design/methodology/approach – A synthetic data set was constructed to prototype proposed data mining techniques and two analyses approaches were investigated. Findings – The two methods were found to yield insights into the problem of the promotion mix in the context of the healthcare industry. First, a factor analysis followed by a regression analysis and an optimization algorithm applied to the resulting equation were used. Second, DAG was used to unravel direct and indirect effects of promotional expenditures on new prescriptions. Research limitations/implications – The data are synthetic and do not incorporate any time autocorrelations. Practical implications – The promotion mix optimization process is demonstrated from the beginning to the end, and the issue of negative coefficient in promotion mix models are addressed. In addition, a method is proposed to identify direct and indirect effects on new prescriptions. Social implications – A better allocation of promotional expenditures has the potential for reducing the cost of healthcare to consumers. Originality/value – The contributions of the paper are two-fold: for the first time in the literature (to the best of the authors’ knowledge), the authors have undertaken a promotion mix optimization process and have carried it through from the beginning to the end Second, the authors propose the use of DAGs to help unravel the effects of various promotion media on sales volume, notably direct and indirect effects.

We demonstrate on a case study with two competing products at a bank how one can use a Hidden Markov Chain (HMC) to estimate missing information on a competitor's marketing activity. The idea is that given time series with sales volumes for products A and B and marketing expenditures for product A, as well as suitable predictors of sales for products A and B, we can infer at each point in time whether it is likely or not that marketing activities took place for product B. The method is successful in identifying the presence or absence of marketing activity for product B about 84% of the time. We allude to the issue of whether, if one can infer marketing activity about product B from knowledge of marketing activity for product A and of sales volumes of both products, the reverse might be possible and one might be able to impute marketing activity for product A from knowledge of that of product B. This leads to a concept of symmetric imputation of competing marketing activity. The exposition in this paper aims to be accessible and relevant to practitioners.

Vehicle longitudinal dynamics system has the characteristics of being strongly non-linear, time-varying, and multiple-perturbed, so, it is difficult to build the mathematical model accurately. The control algorithms, based on accurate mathematical model, can hardly achieve the ideal effect, but control methods, which merely adopt input/output data (I/O) of a system, provides a solution. In this paper, by means of combing model-free adaptive control (MFAC) and sliding-mode control (SMC), the model-free adaptive sliding mode control (MFASMC) method is proposed. By comparison with feedback-feedforward control method, the MFASMC method can better improve the control effect and anti-disturbance performance. Meanwhile, the stability of MFASMC method was proven mathematically. Besides, the parameters of MFASMC method were optimized using genetic algorithm. Results of simulation and HiL test shows that the MFASMC method has fast response, strong robustness and smooth output. It would be better to apply it to the longitudinal dynamics control of intelligent vehicles.

No abstract