Search results

Bankruptcy prediction has attracted a great deal of research in the data mining/machine learning community, due to its significance in the world of accounting, finance, and investment. This chapter examines the influence of different dimension reduction techniques on decision tree model applied to the bankruptcy prediction problem. The studied techniques are principal component analysis (PCA), sliced inversed regression (SIR), sliced average variance estimation (SAVE), and factor analysis (FA). To focus on the impact of the dimension reduction techniques, we chose only to use decision tree as our predictive model and “undersampling” as the solution to the issue of data imbalance. Our computation shows that the choice of dimension reduction technique greatly affects the performances of predictive models and that one could use dimension reduction techniques to improve the predictive power of the decision tree model. Also, in this study, we propose a method to estimate the true dimension of the data.

This research paper aims to investigate detailed relationships between market selection and product positioning decisions and their associated short- and long-term product performance outcomes in the context of the music category: a cultural goods industry with high amounts of product introductions. Market selection decisions are defined by the size, competitiveness and age of market subcategories within an overall product category. Positioning decisions include where a product’s attributes are located spatially in the category (periphery versus the market center), whether a product resides within a single subcategory or spans multiple ones and what brand strategy (single versus co-branding) is used.

Data are from multiple sources for the US music industry (aka product category) from 1958 to 2019 to empirically test the hypotheses: genres (rock, blues, etc.) correspond to subcategories; artists to brands; and songs to products. Regression analyses are used.

A complex set of nuanced results are generated and reported, finding that key marketing decisions drive short-term new product success differently and frequently in opposing ways than long-term success. Launching into very new, well-established or very competitive markets leads to the strongest long-term success, despite less attractive short-run prospects. Positioning a product away from the market center and spanning subcategories similarly poses short-run challenges, but long-run returns. Brand collaborations have reverse effects. Short-run product success is found, overall, to be difficult to predict even with strong data inputs, which has substantial implications for how firms should manage portfolios of products in cultural goods industries. Long-run product success is considerably more predictable after short-run success is observed and accounted for.

While managers and firms in cultural goods industries have long relied on intuition to manage market selection and product positioning decisions, this research tests the hypothesis that objective data inputs and empirical modeling can better predict short- and long-run success of launched products. Specific insights on which song characteristics may be associated with success are found – as are more generalizable, industry-level results. In addition, by distinguishing between short- and long-run success, a more complete picture on how key decisions holistically affect product performance emerges. Many market selection and product positioning decisions have differential impacts across these two frames of reference.