Anomaly detection for indoor air quality (IAQ) data has become an important area of research as the quality of air is closely related to human health and well-being. However, traditional statistics and shallow machine learning-based approaches in anomaly detection in the IAQ area could not detect anomalies involving the observation of correlations across several data points (i.e., often referred to as long-term dependencies). We propose a hybrid deep learning model that combines LSTM with Autoencoder for anomaly detection tasks in IAQ to address this issue. In our approach, the LSTM network is comprised of multiple LSTM cells that work with each other to learn the long-term dependencies of the data in a time-series sequence. Autoencoder identifies the optimal threshold based on the reconstruction loss rates evaluated on every data across all time-series sequences. Our experimental results, based on the Dunedin CO ₂ time-series dataset obtained through a real-world deployment of the schools in New Zealand, demonstrate a very high and robust accuracy rate (99.50%) that outperforms other similar models.

Objective To assess whether non-polluting, more effective home heating (heat pump, wood pellet burner, flued gas) has a positive effect on the health of children with asthma.Design Randomised controlled trial.Setting Households in five communities in New Zealand.Participants 409 children aged 6-12 years with doctor diagnosed asthma.Interventions Installation of a non-polluting, more effective home heater before winter. The control group received a replacement heater at the end of the trial.Main outcome measures The primary outcome was change in lung function (peak expiratory flow rate and forced expiratory volume in one second, FEV1). Secondary outcomes were child reported respiratory tract symptoms and daily use of preventer and reliever drugs. At the end of winter 2005 (baseline) and winter 2006 (follow-up) parents reported their child’s general health, use of health services, overall respiratory health, and housing conditions. Nitrogen dioxide levels were measured monthly for four months and temperatures in the living room and child’s bedroom were recorded hourly.Results Improvements in lung function were not significant (difference in mean FEV1 130.7 ml, 95% confidence interval −20.3 to 281.7). Compared with children in the control group, however, children in the intervention group had 1.80 fewer days off school (95% confidence interval 0.11 to 3.13), 0.40 fewer visits to a doctor for asthma (0.11 to 0.62), and 0.25 fewer visits to a pharmacist for asthma (0.09 to 0.32). Children in the intervention group also had fewer reports of poor health (adjusted odds ratio 0.48, 95% confidence interval 0.31 to 0.74), less sleep disturbed by wheezing (0.55, 0.35 to 0.85), less dry cough at night (0.52, 0.32 to 0.83), and reduced scores for lower respiratory tract symptoms (0.77, 0.73 to 0.81) than children in the control group. The intervention was associated with a mean temperature rise in the living room of 1.10°C (95% confidence interval 0.54°C to 1.64°C) and in the child’s bedroom of 0.57°C (0.05°C to 1.08°C). Lower levels of nitrogen dioxide were measured in the living rooms of the intervention households than in those of the control households (geometric mean 8.5 μg/m3 v 15.7 μg/m3, P<0.001). A similar effect was found in the children’s bedrooms (7.3 μg/m3 v 10.9 μg/m3, P<0.001).Conclusion Installing non-polluting, more effective heating in the homes of children with asthma did not significantly improve lung function but did significantly reduce symptoms of asthma, days off school, healthcare utilisation, and visits to a pharmacist.Trial registration Clinical Trials NCT00489762.