Chih Ming

4. Experimental Results and Discussion 4.1. Statistical Evaluation The NER-SA model’s training set accuracy, test set accuracy, root mean square error (RMSE), and R-square values are used to evaluate the experimental results. We also used the test samples of each dataset to carry out a paired t-test [32] on the actual and predicted values. The statistical study determined by all t values is significant (p < 0.001), implying that the NER-SA model is reliable and feasible.

The importance of (early) parental investments in children’s cognitive and noncognitive outcomes is a question of deep policy significance. However, because parental investments are arguably endogenous, it is a great challenge to empirically estimate their importance. This paper exploits a rich and novel dataset, the China Family Panel Studies, and proposes a culture-specific instrumental variable based on the Chinese zodiac, in order to address this empirical challenge.