Table 3.

Summary of approaches used to predict risk of preterm deliveries.

References	Summary	Key contribution	Dataset used
Fergus et al. (25)	Use of Electrohysterography (the analysis of uterine electrical signals) for diagnosing actual labor and predicting premature birth	Unlike previous works in this domain that focus only on detecting true labor using EHG near the days of delivery, this study uses EHG to even predict term and preterm delivery in early pregnancy	Term-Preterm EHG containing 300 records (38 preterm and 262 term)
Hussain et al. (26)	EHG signals are used to detect preterm births with a novel algorithm	The authors describe a unique dynamic self-organized network immune algorithm for categorizing term and preterm records. The article focuses on boosting sensitivity rates, as forecasting preterm delivery is more crucial than misclassifying a term pregnancy	Term-Preterm EHG
Fergus et al. (27)	Proposed a novel self-organized network immune algorithm that classifies term and preterm records	New electromyography features and feature ranking approaches were used to assess their discriminative powers in detecting term and preterm pregnancies. A comparison of seven different neural networks is performed	Term-Preterm EHG
Despotovic et al. (28)	This study investigates the feasibility of predicting preterm birth from EHG recordings made between the 22nd and 25th week of pregnancy	EHG signals based preterm birth prediction using novel features utilising signal's non-stationarity	Term-Preterm EHG
Gao et al. (29)	Deep learning techniques based Extreme preterm delivery(EPD i.e before the 28th week of pregnancy) prediction	Showed that deep learning algorithms could predict extreme preterm birth (EPB) with the help of temporal relationships in electronic health records (EHRs)	Electronic health records
Jehan et al. (30)	Predicting preterm deliveries using the proteomic and metabolomic characteristics	Established a link between omics data and the prediction of preterm deliveries. Provided a method to predict preterm deliveries in early pregnancy (median gestational age of 13.6 weeks as determined by ultrasonography). PTB prediction accuracy was increased by the use of different omics data sets, implying that PTB is a condition that presents in a variety of biological systems	Blood and urine samples collected from 81 pregnant women. The data was examined from December 2018 to July 2019