Building recommender systems in DataRobot
One of the strengths of driverless artificial intelligence (AI) platforms such as DataRobot lies in their simplification of the data science model-building process. Given the similarity of the content-based recommendation model-building process to the typical ML one, DataRobot's ML capabilities could be leveraged in building these systems. Having set up the data (as detailed in the previous section), click on Start (Figure 10.2) to commence the modeling process. To avoid over-optimistic model performance which fails to generalize where users provide more than one ratings for items, it might be useful to partition the rating according to the users. To do this, within the Advanced options window, open the Group tab and enter user_id in the Group ID Feature field.
As detailed in Chapter 6, Model Building with DataRobot, DataRobot commences the development of ML models when the Start button is clicked. However, with recommender systems...
