Research Article

A Novel Methodology for Credit Spread Prediction: Depth-Gated Recurrent Neural Network with Self-Attention Mechanism

Table 4

Results of the prediction models (loss function = “MAE”)

True valueDJANETDGRUDLSTMDRNNSVRMLPRFVAR

9/2017112.89114.51#109.59118.56115.36108.56116.97117.694121.44
10/2017107.45106.21107.89#113.95112.15106.05113.28112.76115.04
11/2017104.42105.73105.31#110.77108.05101.11106.83109.22112.30
12/2017100.40100.97#102.42107.45105.8397.19105.64105.71111.58
1/201894.7194.0398.69101.68100.3294.72#101.68100.51106.28
2/201896.5295.70#103.21106.88103.2299.7898.77111.14103.21
3/2018109.66104.35111.59116.20111.28110.32#110.56112.88114.01
4/2018112.47112.63#118.29123.56116.87117.08115.34124.02116.36
5/2018115.95116.33#110.80113.83110.56113.03117.52111.91122.35
6/2018123.95122.61117.84121.34118.13117.28124.91#116.67131.73
7/2018122.95124.95#117.51119.97117.96117.32127.17117.46131.63
8/2018117.78116.48#113.22115.88115.23112.03122.08115.09127.66
9/2018115.94117.47112.53113.81113.92109.91121.54116.97#125.37
10/2018117.09115.53113.99113.60113.21110.01116.28#123.65122.99
11/2018132.20130.91#125.84130.23125.76124.81130.04130.63135.14
12/2018151.47150.50#149.44158.19146.64150.15149.32152.88152.51
1/2019150.14151.70150.50#159.20148.31148.73147.88149.66154.39