A clustering model for time-series forecasting

Čorić, Rebeka; Đumić, Mateja; Jelić, Slobodan

doi:10.23919/mipro.2019.8756806

Cited by 1 publication

(2 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence, we experimentally evaluated and demonstrated that our approach outperforms the pollen calendar method, which is typically used in modern pollen prediction systems. In addition, we show that our method outperforms pollen predictions based on patterns [5] and the naive approach that predicts concentration for a given day by copying the pollen concentration seen on the previous day.…”

Section: Contributionmentioning

confidence: 89%

“…Pollen concentrations were collected using the 7-day volumetric For the experiments, the following meteorological data was used: minimal daily temperature (MNT), maximal daily temperature (MKT), precipitation (PAD), relative humidity (VLZ), and maximum daily wind speed (MBV). For every of the three pollen types, all three models (PollenNet, RM, and SM strategy) described in section 2 were used, and the obtained results are compared to the pollen calendar, the naive model, and the prediction based on patterns ( [5]).…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

PollenNet - a deep learning approach to predicting airborne pollen concentrations

Čorić

Matijević

Marković

2023

Croat. oper. res. rev. (Online)

View full text Add to dashboard Cite

The accurate short-term forecasting of daily airborne pollen concentrations is of great importance in public health. Various machine learning and statistical techniques have been employed to predict these concentrations. In this paper, an RNN-based method called PollenNet is introduced, which is capable of predicting the average daily pollen concentrations for three types of pollen: ragweed (Ambrosia), birch (Betula), and grass (Poaceae). Moreover, two strategies incorporating measurement errors during the training phase are introduced, making the method more robust. The data for experiments were obtained from the RealForAll project, where pollen concentrations were gathered using a Hirst-type 7-day volumetric spore trap.Additionally, five types of meteorological data were utilized as input variables. The results of our study demonstrate that the proposed method outperforms standard models typically used for predicting pollen concentrations, specifically the pollen calendar method, pollen predictions based on patterns, and the naive approach.

show abstract

Section: Contributionmentioning

confidence: 89%

Section: Methodsmentioning

confidence: 99%