on 04-Jul-2022 (Mon)

This is not a philosophy book. We remain agnostic about metaphysical concepts like causality and cause. Instead, we focus on the identification and estimation of causal effects in populations, that is, numerical quantities that measure changes in the distribution of an outcome under different interventions.

Part 3 (chapters 14–16) focuses on “the after”: specifically, considerations related to streamlining production release, retraining, monitoring, and attribution for a project. With examples focused on A/B testing, feature stores, and a passive retraining system, you’ll be shown how to implement systems and architectures that can ensure that you’re building the minimally complex solution to solve a business problem with ML

Disadvantages of survey-based CX systems

4. Unfocused: As one executive at a large financial-services company put it, “The association between survey-based scores and business outcomes is not well understood, and, as a result, many parts of the organization simply claim a business impact from their CX initiatives with no real evidence.”

we focus on the identification and estimation of causal effects in populations, that is, numerical quantities that measure changes in the distribution of an outcome under different interventions.

agent-based models are often not developed explicitly for prediction, and are generally not validated as such. We therefore present a novel data-driven agent-based modeling framework, in which individual behavior model is learned by machine learning techniques, deployed in multi-agent systems and validated using a holdout sequence of collective adoption decisions.

the California Solar Initiative (CSI), a rooftop solar subsidy program initiated in 2007 with the goal of creating 1940 megawatts of solar capacity by 2016 [12]. The CSI program has been touted as a great success, and it certainly seems so: over 2000 megawatts have been installed to date. However, in a rigorous sense, success would have to be measured in comparison to some baseline; for ex- ample, in comparison to the same world, but without incentives. Of course, such an experiment is impossible in practice. However, in principle, insight can be drawn by sensitivity analysis based on hypothetical solar diffusion model. What is the most appropriate modeling methodology to build a highly robust solar diffusion mode

Agent-based modeling (ABM) has long been a common framework of choice for studying aggregate, or emergent, properties of complex systems as they arise from microbehaviors of a multitude of agents in social and economic contexts [7,30,35]. ABM appears well-suited to policy experimentation of just the kind needed for the rooftop solar market. Indeed, there have been several attempts to develop agent-based models of solar adoption trends [14,32, 38]. Both traditional agent-based modeling, as well as the specific models developed for solar adoption, use data to calibrate aspects of the models (for example, features of the social network, such as density, are made to match real networks), but not the entire model. More importantly, validation is of- ten qualitative, or, if quantitative, using the same data as used for calibration. The weakness of validation makes those models less eligible as a reliable policy experiment tool.