The elastic network regression between iron nutrition and thyroid hormone
The elastic network regression was used to further explore of the specific relationship between iron nutrition and thyroid function. Based on the results of canonical correlation and Spearman correlation, we chose FT3 as the dependent variable, whereas other factors were the independent variables. In our study, all factors were incorporated into the elastic network regression forecasting model (Alpha = 0.1, Lambda = 0.0032) (Fig. 3 ), and the equations could be established as follows.
FT3 = 0.3802 + 0.7191T3 – 0.0022T4 + 0.1062FT4 – 0.0014TSH + 0.0014SF + 0.0077Hb
From the results of elastic network regression, we found that FT3 was influenced by T3, and the FT4 was the second; these findings were consistent with the simple correlation results in our study. However, they were inconsistent with the canonical correlation results. Thus, based on the above results, it was still impossible to get an exact result to prove whether SF or HB played a more important role in the thyroid function. It was also proved that iron nutrition affected thyroid function in many ways.