A Bayesian Approach to Ranking Private Companies Based on Predictive Indicators

Private equity investors seek to identify potential investment opportunities in growth stage private companies and rank their prospects relative to a cohort of companies such as an industry sector. Growth stage private companies often have investment transaction histories from which industry specific characteristics associated with successful and failed companies may be discerned using machine learning. In general, one of the primary challenges in pursuing this approach is the sparsity of historical data on private companies which is exacerbated in nascent sectors by the relatively few number of observed exits. This papers describes a four step predictive approach for ranking private companies within a cohort which can be applied to sparse industry specific historical data. This papers describes a four step approach based on (i) extracting and selecting features; (ii) training Support Vector Machine (SVM) classification models from feature pairs of labeled companies in an industry; (iii) estimating posterior probabilities of success and failure given a set of SVM model outputs; and (iv) ranking unlabeled companies within a cohort based on scores derived from posterior probability estimates. The main advantage of this ranking approach is that it includes labeled companies with missing features which would otherwise be excluded if the approach was based on the output of a single SVM model trained from higher dimensional feature sets. Being able to include labeled companies with missing data is a critical step towards a machine learning based methodology for ranking companies relative to an industry sector and we anticipate that this approach will not only be of interest to machine learning specialists with an interest in venture capital and private equity but extend to a broader readership whose interest is in other classification models applied to finance where missing data is the primary obstacle