Next Article in Journal
Resource-Efficient Parallelized Random Access for Reliable Connection Establishment in Cellular IoT Networks
Next Article in Special Issue
Multispectral Plant Disease Detection with Vision Transformer–Convolutional Neural Network Hybrid Approaches
Previous Article in Journal
Detection and Mitigation of SYN Flooding Attacks through SYN/ACK Packets and Black/White Lists
Previous Article in Special Issue
An Optimized Ensemble Deep Learning Model for Predicting Plant miRNA–IncRNA Based on Artificial Gorilla Troops Algorithm
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sensors
Volume 23
Issue 8
10.3390/s23083818
sensors-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Predicting Daily Aerobiological Risk Level of Potato Late Blight Using C5.0 and Random Forest Algorithms under Field Conditions

by
Laura Meno
1,
Olga Escuredo
1,*,
Isaac K. Abuley
2 and
M. Carmen Seijo
1
1
Department of Vegetal Biology and Soil Sciences, Faculty of Sciences, University of Vigo, 32004 Ourense, Spain
2
Department of Agroecology, Flakkebjerg Research Center, Aarhus University, Forsøgsvej 1, 4200 Aarhus, Denmark
*
Author to whom correspondence should be addressed.
Sensors 2023, 23(8), 3818; https://doi.org/10.3390/s23083818
Submission received: 25 February 2023 / Revised: 3 April 2023 / Accepted: 6 April 2023 / Published: 8 April 2023
(This article belongs to the Special Issue Precision Agriculture with Deep and Machine Learning)
Browse Figures
Versions Notes

Abstract

:
Late blight, caused by Phytophthora infestans, is a major disease of the potato crop with a strong negative impact on tuber yield and tuber quality. The control of late blight in conventional potato production systems is often through weekly application of prophylactic fungicides, moving away from a sustainable production system. In support of integrated pest management practices, machine learning algorithms were proposed as tools to forecast aerobiological risk level (ARL) of Phytophthora infestans (>10 sporangia/m3) as inoculum to new infections. For this, meteorological and aerobiological data were monitored during five potato crop seasons in Galicia (northwest Spain). Mild temperatures (T) and high relative humidity (RH) were predominant during the foliar development (FD), coinciding with higher presence of sporangia in this phenological stage. The infection pressure (IP), wind, escape or leaf wetness (LW) of the same day also were significantly correlated with sporangia according to Spearman’s correlation test. ML algorithms such as random forest (RF) and C5.0 decision tree (C5.0) were successfully used to predict daily sporangia levels, with an accuracy of the models of 87% and 85%, respectively. Currently, existing late blight forecasting systems assume a constant presence of critical inoculum. Therefore, ML algorithms offer the possibility of predicting critical levels of Phytophthora infestans concentration. The inclusion of this type of information in forecasting systems would increase the exactitude in the estimation of the sporangia of this potato pathogen.

1. Introduction

Late blight, caused by Phytophthora infestans, is a major disease of potato crop, with a strong negative impact on tuber yield and quality [1,2,3]. The pathogen is regarded as a threat to global food security because worldwide losses due to late blight are estimated to exceed annually $5 billion [4,5]. P. infestans is described as a lower water oomycete and infects the potato crop through the tuber and soil during cool and wet weather. Infection of shoots can be caused by mycelium growing from the tuber into the developing shoot or through sporangia and zoospores formed on the tuber surface under wet conditions [6]. Then, the potential risk of disease development depends in part on the aerial transport of P. infestans sporangia to potato fields from neighbor’s infection fields [7,8,9,10]. Depending on host susceptibility and environmental conditions, the first symptoms can be visible 3–4 days after infection [1]. Night temperatures of 10–16 °C accompanied by light rain, fog or heavy dew and followed by days of 13–16 °C with high relative humidity are ideal conditions for late blight infection and development [11]. The first symptoms are followed by the production of new sporangia and the infection cycle is repeated as many times as the weather conditions allow the viability of the released sporangia [6,8,12,13].
Potato crops contributed to alleviating world hunger and potatoes remain one of the agricultural resources needed in line with the zero-hunger sustainable development goal. However, the crop management strategies increasingly require the application of phytochemicals, which may cause undesirable effects on the environment and on human health, with an increasing production costs for growers. For effective and ecofriendly management of potato late blight, scientific and technical efforts must be made to understand how the disease progresses and how this progression can be slowed down. One of the important issues in late blight management is to forecast when, where and how abundant airborne inoculum will be, to prevent the onset of the epidemic. The airborne inoculum of the pathogen appears to have a significant impact on the disease epidemic. However, the prediction of airborne spores of plant pathogens is difficult because they are influenced by a plethora of factors (temperature, relative humidity, leaf wetness, wind, phenological stage) [7,10,12,14,15]. Hence, the efforts towards understanding and predicting airborne sporangia of P. infestans based on multiple factors that condition its development, with a significant impact on the management of late blight, with less fungicide applications are sought.
In recent years, the agricultural sector was able to adopt the main technological innovations relying on artificial intelligence (AI), artificial neural networks (NN) and machine learning (ML). The goal is to digitize itself and increase the autonomy of many processes by making better data-driven decisions, reducing the workload, inputs and increase the quality of the final product [16,17,18,19,20]. The multi-view spectral information from unmanned aerial vehicles (UAV) based color-infrared images combined with machine learning algorithms was used to improve the estimation of nitrogen nutrition status in winter wheat and optimize the fertilization [17]. Classification methods and clustering trough image analyses such as neural networks (CNN) were used to simulate the humans’ decision-making process. CNNs were shown to have great potential for fine classification problems using an image of the same object from different views [18,19]. Decision trees, support vector machines or k-means together with information from foliage of the crop were used in precision agriculture and the effective detection, identification and quantification of plant diseases [21,22]. In the case of potato crop, ML algorithms were recently applied for monitoring diseases through image-based techniques [20,23,24,25,26,27]. Sugiura et al. [24] proposed a phenotyping system for mapping late blight on potato crop by analyzing pixel change between consecutive images. The assessment of late blight severity in potato by acquiring high resolution multispectral images with a low-cost camera and ML algorithms was also reported [25]. More recently, the early detection and severity assessment of late blight in potato crops by multispectral imagine were evaluated [26,27]. However, these studies focused on the detection and identification of disease after the onset of the infection process. The preceding step is the early detection of inoculum in the environment of the crop able to cause first late blight symptoms on the potato canopy. Aerobiology is an excellent discipline for this purpose, allowing real time knowledge of sporangia in the potato atmosphere [9,10,14,15,28,29,30].
There were some aerobiological studies that focused on understanding the influence of climatic factors on the dynamics of spores in the atmosphere of the potato crop using different multivariate statistical techniques and ML algorithms [31,32,33]. However, there were fewer studies trying to predict P. infestans sporangia levels in the environment crop [10,15]. Furthermore, despite the great available scientific information on late blight, few studies focused on the specific value of airborne sporangia concentration as a monitoring tool for late blight control [10,14,15,29,34,35]. In this sense, with the purpose of estimating the late blight risk during the early stages of potato crop development, ML algorithms were applied. The goals of the present study were: (i) to assess the concentration of P. infestans in each phenological stage of potato crop in northwest Spain; (ii) to derive a simple binary classification model for predicting the days exceeding the aerobiological risk level of pathogen; and (iii) to validate ML algorithms as a tool for forecasting late blight outbreaks.

2. Materials and Methods

2.1. General Aspects of the Experimental Potato Field

The experimental field was located in A Limia (Galicia, northwest Spain) and the study was conducted over five crop seasons (2017–2021). A 4-hectare field was planted with the potato cultivar Agria, considered as a medium resistant to late blight disease. The field plot was managed under an annual wheat–potato rotation system. The dates of planting and the main phenological stages for each season are shown in Table 1. The phenological monitoring started when 50% of plants emerged, and weekly observations until crop senescence were performed. Three main phenological stages considering the BBCH scale of Hack et al. [36] with some modifications [37] were monitored: foliar development (FD), flowering (FL), senescence (SE).

2.2. Weather Monitoring

Weather data were registered hourly using a portable weather station i-METOS 3.3. (Pessl Instruments, Weiz, Austria) placed in the middle of the experimental potato field, at 1.5 m height since 50% of plants were emerged. Daily mean temperature (T, °C), relative humidity (RH, %), leaf wetness (LW, h) and wind speed (Wind, m/s) were calculated with hourly data registered.
The release of sporangia from the sporangiophore (expressed as spore release, SR) was calculated using hourly RH and a critical value of 88% of RH [7]. The escape of sporangia from the canopy into the atmosphere (escape) was calculated with the hourly wind speed data according to the formula proposed by Skelsey et al. [7].
Following the Danish late blight model (BlightManager), the daily risk value (DRV) based on hourly temperature and relative humidity was calculated [38]. Subsequently, the infection pressure (IP), which is a running sum of the DRV of five days, was calculated as described previously [10,38].

2.3. Aerobiological Sampling

The surveillance of the atmosphere of the crop was performed through an aerobiological sampler type Lanzoni VPPS 2000 (Lanzoni S.r.l., Bologna, Italy) placed in the potato plot at 1.5 m high close to the weather station since the 50% of plants emerged. The equipment contained a vacuum pump to aspirate the airborne particles surrounding through a horizontal aperture, which was connected with the melinex tape covered where these particles were retained. A clockwork-driven drum containing the tape was moving continuously for 7 days, and each week, the tape was replaced. The methodology used for the assembly and counting of airborne particles was based on the proposal by Galán et al. [39], expressing the concentration of P. infestans in sporangia/m3.
Based on field experiences, a daily aerobiological risk level (ARL) of 10 sporangia/m3 was established. This level defines the concentration of sporangia likely to cause late blight infection in the potato crop, considered in the statistical treatment of ML algorithms (Figure 1).

2.4. Data Analyses

Data curation and statistical analyses were performed with the R language and environment for statistical computing version 4.1.3 [40] and IBM SPSS Statistics 22 program. Spearman correlation test was applied to analyze the relationships between weather factors and presence of sporangia until seven previous days, with a significance level of p < 0.05.
Prior to the treatment of ML algorithms, a binary code based on the ARL established by the aerobiological method explained above was proposed. The code established the zero (0) value for days with a concentration lower than 10 sporangia/m3, while the days that exceeded this concentration threshold were classified as 1 (Figure 1). Then, ML algorithms C5.0 and random forest (RF) to predict ARL were applied. The C5.0 algorithm is a decision tree ML algorithm that uses entropy (a measure of the randomness in a partition) for splitting trees. The C5.0 model uses adaptive boosting to improve the predictive power of the final model. The RF algorithms, also known as decision tree ensemble algorithms, combine the results of multiple independent decision trees to make predictions about new data sets [41]. Each tree in the forest assigns the most probable class label to each input.

Implementing the Machine Learning Algorithms

The data set was split into 80% (469 values/rows) training and 20% (117 values/rows) test data sets. As the tested algorithms were not based on distance metrics, no standardization was carried for the data. These algorithms were implemented with “train” function in the caret R package version 6.0-92 [42]. The method option in “train” function was set to “C5.0” and “rf” for implementing the C5.0 and RF algorithms, respectively. For all algorithms, 10-fold cross-validation (CV) was used to optimize the models. For the C5.0, both the tree and rule-based models with or without winnowing (i.e., a process of removing uninformative predictors) were evaluated and the best model was selected. The hyperparameter (the node size) in the RF algorithm by comparing the accuracy of models from different node sizes (1 to 10) was also optimized. The node size that resulted in the highest accuracy was selected for building the RF model.
The following metrics were used to evaluate the models: (a) accuracy (the percentage of correct predictions by the model), and kappa statistics/accuracy (an adjustment to predictive accuracy by accounting for the possibility of a correct prediction by chance alone). These metrics were computed from the “confusionMatrix” function in the caret R package version 6.0-92 [42]. In order to evaluate the performance of the proposed models, sensitivity and specificity were considered. Sensitivity refers to a proportion of days with sporangia concentrations >10 that correctly give, a positive result using the test in question. Specificity means the proportion of days with a concentration <10 sporangia/m3 that correctly give a negative result using the test in question.
Model performance was also assessed using receiver operating characteristics (ROC). A ROC curve is a graphical representation of sensitivity versus specificity for a binary classifying system. Thus, in ROC analysis, the area under the curve (AUC) represents the power of a parameter to discriminate between two classes. ROCs are used to evaluate the trade-off between true and false positives. A ROC curve shows the model’s prediction based on the positive class when the actual result is positive. Thus, in ROC analysis, the area under the curve represents the power of a parameter to discriminate between two classes. AUC values are interpreted as follows: <0.6 (no discrimination), 0.6–0.7 (poor), 0.7–0.8 (acceptable), 0.8–0.9 (excellent), >0.9 (outstanding). The ROC and AUC was computed with the ROCR package (version 1.0-11) [43].

3. Results

3.1. Crop Phenology during Period of Study

Planting dates in the five years took place between April and May (Table 1). The earliest planting year was 2017, while the latest planting year was 2020. Emergence occurred at 15 days, except in 2017, when it was delayed by 25 days. Flowering occurred in the 34–40 days after emerging, except in 2017, when it happened a few days earlier. Considering the days that elapsed since emergence, flowering occurred between days 29 (2017) and 40 (2019). The longer crop cycle (2018) reached the end of senescence at 106 days after emergence and the last two crop cycles, 2021 and 2020, were the shortest cycles and reached senescence at 78 and 83 days after emergence, respectively.

3.2. Overview of Weather Conditions by Phenological Stage

Specific mean values by phenological stage are presented in Table 2. Of the five years studied, the 2018 growing season was the wettest, with approximately 297 mm of water on 37 rainy days. In contrast, 2020 was the warmest and driest year. During the whole cycle, only 15 mm of rainfall was recorded in 7 days. In general, the foliar development (FD) was the stage with the mildest temperatures and the rainiest days, although not the most abundant in terms of rain amount. The average temperature in this phenological stage was 16.4 °C, with a mean of 10.6 days of rain. Additionally, this phenological stage was the one with the highest RH: 78.1% on average. On the contrary, during the senescence stage (SE), the lowest RH was recorded, with an average of 72.6%. In addition, in this final period, there were fewer rainy days and, consequently, less accumulated water (a mean of 4 days and 29.0 mm). The flowering (FL) was the stage with the highest average temperatures (19.4 °C) and mean RH (73.7%).

3.3. Daily Sporangia Concentration by Main Phenological Stage

The most abundant seasons in terms of daily sporangia concentration were 2018 and 2021, with 2271 and 1836 accumulated sporangia in each crop cycle, respectively (Table 3). Conversely, the year that stood out for its low concentration of sporangia was 2020, with a total of 36 sporangia in the whole cycle. In addition, in 2020, days with ARL were not found. In terms of phenological phase, the foliar development stage (FD) was the stage in which the highest daily sporangia concentrations were recorded. Additionally, the highest number of days with ARL was recorded during the foliar development stage (FD) except in 2017. In 2019, 100% of days with ARL was recorded during FD stage (Table 3).

3.4. Relationships between the Sporangia Concentration and the Meteorological Parameters

Spearman correlation coefficients were calculated between the meteorological variables and sporangia concentration. The highest positive and significant coefficients were found between the presence of sporangia of one and two previous days with daily sporangia concentration. Significantly negative coefficients between T and SR and the daily sporangia concentration were found (p < 0.05), while significantly positive coefficients between the sporangia and RH and IP were found. Wind and escape were positively correlated with sporangia from the second previous day until the fifth previous day. For LW, the correlation between sporangia concentration and the same-day recording (LW_0) was significantly positive, in contrast to the previous days (Table 4).

3.5. Machine Learning Algorithms to Predict the Daily Sporangia Risk Level

In order to develop a model to predict the ARL, ML algorithms (RF and C5.0) were applied using airborne daily sporangia and weather data. A binary prediction of 1 and 0 were designed (1: corresponded with days in which sporangia concentrations were equal or higher than 10 sporangia/m3; 0: days with concentrations lower than this value). Depending on the model, the ranking of the importance of the variables used for the developed model varied (Table 5). The C5.0 algorithm ranked the variables according to the usage ratio of each variable in the final result. The most decisive variables in the prediction performed by the C5.0 algorithm were Wind (_6, _2), the sporangia (_7, _1), T_7, LW_0, IP_0, SR_5 and RH (_7, _0). According to the RF algorithm, sporangia was the most influenced variable on the accuracy of model followed by SR_6, Escape_2 and IP_0. The sporangia of previous days (Sporangia_1) was ranked as the most important variable with a 36% decrease in accuracy.
Aerobiological variables related to the presence of sporangia on previous days were the most important variables in the prediction performed by the RF algorithm, coinciding with the higher significant Spearman correlation coefficients (p < 0.01) (Table 4). However, the weather variables selected (SR_6, Escape_2, IP_0) for the prediction by this algorithm were those with the highest significant Spearman correlation coefficients (p < 0.01). Most of the weather variables selected by the C5.0 algorithm (T_7, Wind_2, LW_0, IP_0) also had the highest significant correlation coefficients (p < 0.01) within their weather category.
The accuracy of two ML algorithms applied is shown in Table 6. The suitable node size was 1 for the RF model. For the C5.0 model, a tree-based model without winnowing was the best model. The results of prediction showed an accuracy of 87% and 85% for RF and C5.0, respectively. The Kappa values were 0.70 for RF and 0.65 for C5.0, which indicated that the RF results were less conditioned by chance than those provided by C5.0 algorithm. In both cases, the results were acceptable. Considering the parameter that indicates the percentage of success in the prediction of days with a sporangia concentration equal to or greater than 10 (Sensitivity), the C5.0 model was more accurate. On the contrary, the prediction of the models with the daily concentration of sporangia lower than the established threshold (specificity), RF model had a higher prediction of 0.92. However, despite using different relevance variables in the classification algorithms, the sporangia prediction accuracies were similar.
For the choice of the best algorithm, ROC curves were used to compare the ratio of false positives versus true positives (Figure 2). Both algorithms presented high and similar areas under the curve (AUC), 0.904 and 0.903 for the algorithm C5.0 and RF, respectively.

3.6. Checking Aerobiological Risk Level Prediction by Machine Learning Algorithms

The resulting predictions from each model algorithm with the observed SRL (aerobiological data) were compared. The prediction result of applying each of the algorithms to the sporangia level from the five growing seasons based on the SRL is shown in Figure 3. Overall, the algorithms made an accurate prediction in all five years of the study, even in the anomalous year of 2020. In this year, no day with a daily concentration of more than 10 sporangia/m3 was recorded and the algorithms did not predict risk for that year. The season with the highest number of days with a daily concentration above 10 sporangia/m3 were 2018 and 2021. In both high-risk and low-risk years, the two algorithms showed the prediction in agreement with the observed data (Figure 3).
Most of the positive predictions were shown in the first months of crop, coinciding with the phenological stages of FD and FL, from May to the beginning of July, except for in the year 2021, where the possibility of infection was extended to the beginning of August. In 2017, the C5.0 algorithm predicted the first risk day at the end of May a day in advance.

4. Discussion

The understanding of different factors that support the aerial dispersal of P. infestans for the correct prediction of late blight epidemics is crucial [7,8,10,14,15,34,44]. However, experimental studies under field conditions with the aim of predicting sporangia are limited. In this sense, the present study is one of the few that predicted airborne sporangia using the value of the inoculum quantified in the crop environment and at the same time, considering climatic factors. Several factors (e.g., wind, temperature, solar radiation, rain) could be used to forecast the outbreak of late blight [45]. This study focused on the main weather factors (T, RH, LW and wind) and derived variables (IP, SR and escape) with inoculum quantity to predict daily sporangia risk level that cause outbreaks of potato late blight. The results of this research support the importance of considering inoculum to make decisions of fungicides applications. This information combined with decision support system for late blight would improve the exactitude of warnings for a correct management of fungal treatments. The first application rate could be predicted before the onset of the disease in the field and consequently, the number of chemical applications would decrease.
According to previous research, linear regression models or neural network can predict with success disease behavior such as mycotoxin secretion in crops such as grapevine, rice and wheat or improve the ability of crop-growth monitoring [17,46,47,48,49,50]. In potato, the publications related with prediction of inoculum of late blight in air using ML algorithms are non-existent. For effective management of potato late blight, efforts must be made to slow the progress of the disease, especially by reducing the primary inoculum. In this sense, ML algorithms were applied because of their usefulness in managing large databases with multiples variables, images or spectra [17,18,19]. The performance of the models was evaluated as a binary classification system, categorizing the daily sporangia into two classes by aerobiological criteria (ARL) according to the daily sporangia concentration greater or less than 10 sporangia/m3. The C5.0 and RF algorithms were good and robust algorithms to predict days with high detached sporulation of P. infestans, resulting a ROC of 0.903 and 0.902, respectively. These ML algorithms provided good predictions in other potato pathogen, such as Alternaria spp. [37]. The ranking of the variables of importance by RF and C5.0 algorithms took the sporangia variables from previous days as the most influential to develop the prediction models. This suggests a strong influence of previous sporangia counts in predicting sporangia level at present, as also corroborated by the Spearman correlation test. A strong influence of sporangia of previous days in the presence of sporangia of present day was observed. This fact coincides with previous studies on P. infestans [10,15] and Alternaria [33,37,51,52] on potato crop, as well as Botrytis cinerea [53] and Uncinula necator [49] on vineyards. This trend emphasizes the need to continually monitor the airborne spores, as they will be needed to accurately predict future spread of spores.
To predict the behavior of P. infestans, it is important to know about crop season and geographical area [7]. Typically, the potato growing seasons in northwest Spain run from the beginning of May to the end of September. During this period, the temperature is the key weather factor for the growth of the potatoes. Throughout the growing season, spring rains are common and contribute to the development of the crop during the first few weeks. However, temperatures and high humidity also prove to be suitable for the presence of certain diseases for potato crop, such is the case of late blight. The production, dissemination and germination of P. infestans sporangia, as well as their penetration into host tissues are particularly influenced by mild temperatures and high relative humidity [1,7,8,10,14,15,34,44]. According to the Spearman correlation test, the coefficients between T and the sporangia presence were significantly negative. On the contrary, RH and LW variables showed a significantly positive relationship with the sporangia presence. Both algorithms agreed with IP and SR as the variables of higher importance. It is known that optimum temperatures for late blight epidemics are between 16 and 23 °C [6]. The IP variable combines the effect of T and RH under one variable, considering optimum values to late blight infection of 10–24 °C and RH > 88% [10]. In the studied area, mean temperatures between 16–21 °C were repeated during the whole period. In addition, rains and high humidity in the first months of crop development can explain higher sporangia concentration during foliar development stage. Under optimal temperature range to late blight epidemics, the lack of rain and dry weather could decrease the infection process and sporulation, as our results showed, with a decrease in rainy days and RH in flowering and senescence stages. Furthermore, it was shown that temperatures above 28 °C negatively affect sporangia production [54,55]. Thus, this could explain the lowest sporangia concentration trapped during senescence of five studied crop seasons.
The results agree with previous studies that showed changing dynamics of the P. infestans inoculum concentration during the potato growing season [9,10,14,28,34]. However, these studies focused on the prediction of late blight risk solely based on weather factors, while assuming the constant presence of inoculum [55,56,57]. Although the infection of potato plants with P. infestans is highly dependent on weather conditions, late blight epidemics cannot be explained exclusively by weather data [7,54]. Aerobiological information can be useful to avoid false alarms of infection risk when climatic conditions are favorable, but if there is no presence of inoculum in the air, there is no risk [34]. The variability in the sporangia concentrations during each phenological stage confirm that the presence of aerial inoculum in any place is not an unlimited factor. Thus, the climatic conditions, the local topography, the associated fungal host and the phenological state of the plant condition affect this variation [9,10,14,15,30]. Despite the scarcity of studies that considered quantification of the airborne inoculum, the present results agree with researchers that support its presence as essential to know the real late blight pressure in a particular area and to predict new reinfections [9,10,13,14,15,29,35,54]. However, for the success of the proposed methodology, several key factors are necessary: having several years of study, meteorological stations on the plot itself or in the vicinity of the monitored area, and specialized personnel for the extraction of aerobiological and meteorological data, as well as the correct treatment of algorithms with complex statistical methods. This type of research has great practical utility in the agricultural sector because it allows farmers to detect the infection before it manifests itself. Consequently, these professionals achieve greater effectiveness in the application of chemical treatments, and reduce investment in preventive treatments. The search for more environmentally sustainable agricultural solutions and tools to minimize the impact of these changing fungal diseases in recent years will have an impact on the economic value of the final product placed on the market. At the same time, it will favor food security and human health in the world.

5. Conclusions

The development of a simple and robust tool to forecast days with a high amount of inoculum capable of developing infection of late blight in potato crops can be useful for the sector. The information on the inoculum concentrations of P. infestans in the crop environment through aerobiological sampling was an important variable to consider in the development of ML techniques. The two ML algorithms applied (RF and C5.0) were accurate for the prediction of the aerobiological risk level (ARL) using meteorological parameters and inoculum of the pathogen from previous days. The different statistical treatments highlighted the influence of the sporangia from previous days. Furthermore, the most influential meteorological variables in the two validated ML algorithms were IP and SR. Therefore, it is interesting to integrate multiple meteorological variables and the presence of sporangia in the crop environment to guarantee the success of these prediction systems. This approach of using a classification model to forecast late blight risk may improve the accuracy of disease risk warning systems for potato plants in the study area. Thus, the utility of incorporating the inoculum present in the crop environment with weather variables in ML algorithms for monitoring of plant pathology in sustainable agricultural systems was demonstrated.

Author Contributions

Conceptualization, L.M., I.K.A., O.E. and M.C.S.; formal analysis, L.M., I.K.A. and O.E.; investigation, L.M. and I.K.A.; data curation, L.M. and I.K.A.; writing—original draft preparation, L.M. and I.K.A.; writing—review and editing, L.M., O.E., I.K.A. and M.C.S.; supervision, I.K.A., O.E. and M.C.S.; project administration, M.C.S.; funding acquisition, M.C.S. All authors have read and agreed to the published version of the manuscript.

Funding

Laura Meno is a beneficiary of the pre-PhD contract FPU 17/00267 granted by the Ministry of Education, Culture, and Sports from Spain.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare that they have no conflict of interest.

References

  1. Cooke, B.M.; Jones, D.G.; Kaye, B. The Epidemiology of Plant Diseases; Springer: Dordrecht, The Netherland, 2006; 568p. [Google Scholar]
  2. Dowley, L.; Grant, J.; Griffin, D. Yield losses caused by late blight (Phytophthora infestans (Mont.) de Bary) in potato crops in Ireland. Ir. J. Agric. Food Res. 2008, 47, 69–78. [Google Scholar]
  3. Filippov, A.; Kuznetsova, M.; Rogozhin, A.; Iakusheva, O.; Demidova, V.; Statsyuk, N. Development and testing of a weather–based model to determine potential yield losses caused by potato late blight and optimize fungicide application. Front. Agric. Sci. Eng. 2018, 5, 462–468. [Google Scholar]
  4. Latijnhouwers, M.; Ligterink, W.; Vleeshouwers, V.G.; van West, P.; Govers, F. A Galpha subunit controls zoospore motility and virulence in the potato late blight pathogen Phytophthora infestans. Mol. Microbiol. 2004, 51, 925–936. [Google Scholar] [CrossRef] [PubMed]
  5. Guenthner, J.; Michael, K.; Nolte, P. The economic impact of potato late blight on US growers. Potato Res. 2001, 44, 121–125. [Google Scholar] [CrossRef]
  6. Harrison, J. Effects of the aerial environment on late blight of potato foliage–A review. Plant Pathol. 1992, 41, 384–416. [Google Scholar] [CrossRef]
  7. Skelsey, P.; Kessel, G.; Holtslag, A.; Moene, A.; Van Der Werf, W. Regional spore dispersal as a factor in disease risk warnings for potato late blight: A proof of concept. Agric. For. Meteorol. 2009, 149, 419–430. [Google Scholar] [CrossRef] [Green Version]
  8. Aylor, D.E.; Schmale III, D.G.; Shields, E.J.; Newcomb, M.; Nappo, C.J. Tracking the potato late blight pathogen in the atmosphere using unmanned aerial vehicles and Lagrangian modeling. Agric. For. Meteorol. 2011, 151, 251–260. [Google Scholar] [CrossRef]
  9. Fall, M.; Van der Heyden, H.; Brodeur, L.; Leclerc, Y.; Moreau, G.; Carisse, O. Spatiotemporal variation in airborne sporangia of Phytophthora infestans: Characterization and initiatives towards improving potato late blight risk estimation. Plant Pathol. 2015, 64, 178–190. [Google Scholar] [CrossRef]
  10. Meno, L.; Abuley, I.; Escuredo, O.; Seijo, M. Factors influencing the airborne sporangia concentration of Phytophthora infestans and its relationship with potato disease severity. Sci. Hortic. 2023, 307, 111520. [Google Scholar] [CrossRef]
  11. Kirk, W.; Wharton, P.; Hammerschmidt, R.; Abu–el Samen, F.; Douches, D. Late Blight. Michigan State University Extension Bulletin E–2945. East Lansing, MI, USA. 2013. Available online: http://www.potatodiseases.org/lateblight.html (accessed on 20 July 2021).
  12. Skelsey, P. Multi–Scale Modeling of Potato Late Blight Epidemics; Wageningen University and Research: Wageningen, The Netherlands, 2008; 258p. [Google Scholar]
  13. Skelsey, P.; Dancey, S.; Preedy, K.; Lees, A.; Cooke, D. Forecasting the spread of aerially transmitted crop diseases with a binary classifier for inoculum survival. Plant Pathol. 2018, 67, 920–928. [Google Scholar] [CrossRef] [Green Version]
  14. Escuredo, O.; Seijo–Rodríguez, A.; Rodríguez–Flores, M.S.; Seijo, M.C. Decision support systems for detecting aerial potato Phytophthora infestans sporangia in northwestern Spain. Agron. J. 2019, 111, 354–361. [Google Scholar] [CrossRef]
  15. Seijo–Rodríguez, A.; Escuredo, O.; Rodríguez–Flores, M.S.; Seijo, M.C. Improving the use of aerobiological and phenoclimatological data to forecast the risk of late blight in a potato crop. Aerobiologia 2018, 34, 315–324. [Google Scholar] [CrossRef]
  16. Belattar, S.; Otman, A.; Khatir, H. Comparing machine learning and deep learning classifiers for enhancing agricultural productivity: Case study in Larache Province, Northern Morocco. Int. J. Electr. Comput. 2023, 13, 1689–1697. [Google Scholar] [CrossRef]
  17. Lu, N.; Wu, Y.; Zheng, H.; Yao, X.; Zhu, Y.; Cao, W.; Cheng, T. An assessment of multi–view spectral information from UAV–based color–infrared images for improved estimation of nitrogen nutrition status in winter wheat. Precis. Agric. 2022, 23, 1653–1674. [Google Scholar] [CrossRef]
  18. Seeland, M.; Mäder, P. Multi–view classification with convolutional neural networks. PLoS ONE 2021, 16, e0245230. [Google Scholar] [CrossRef] [PubMed]
  19. Guarino, A.; Malandrino, D.; Zaccagnino, R.; Capo, C.; Lettieri, N. Touchscreen gestures as images. A transfer learning approach for soft biometric traits recognition. Expert Syst. Appl. 2023, 219, 119614. [Google Scholar] [CrossRef]
  20. Lizarazo, I.; Rodriguez, J.L.; Cristancho, O.; Olaya, F.; Duarte, M.; Prieto, F. Identification of symptoms related to potato Verticillium wilt from UAV–based multispectral imagery using an ensemble of gradient boosting machines. Smart Agric. Technol. 2023, 3, 100138. [Google Scholar] [CrossRef]
  21. Zhang, X.; Han, L.; Dong, Y.; Shi, Y.; Huang, W.; Han, L.; González–Moreno, P.; Ma, H.; Ye, H.; Sobeih, T. A deep learning–based approach for automated yellow rust disease detection from high–resolution hyperspectral UAV images. Remote Sens. 2019, 11, 1554. [Google Scholar] [CrossRef] [Green Version]
  22. Bagheri, N. Application of aerial remote sensing technology for detection of fire blight infected pear trees. Comput. Electron. Agric. 2020, 168, 105147. [Google Scholar] [CrossRef]
  23. Hou, C.; Zhuang, J.; Tang, Y.; He, Y.; Miao, A.; Huang, H.; Luo, S. Recognition of early blight and late blight diseases on potato leaves based on graph cut segmentation. J. Agric. Food Res. 2021, 5, 100154. [Google Scholar] [CrossRef]
  24. Sugiura, R.; Tsuda, S.; Tamiya, S.; Itoh, A.; Nishiwaki, K.; Murakami, N.; Shibuya, Y.; Hirafuji, M.; Nuske, S. Field phenotyping system for the assessment of potato late blight resistance using RGB imagery from an unmanned aerial vehicle. Biosyst. Eng. 2016, 148, 1–10. [Google Scholar] [CrossRef]
  25. Duarte–Carvajalino, J.M.; Alzate, D.F.; Ramirez, A.A.; Santa–Sepulveda, J.D.; Fajardo–Rojas, A.E.; Soto–Suárez, M. Evaluating late blight severity in potato crops using unmanned aerial vehicles and machine learning algorithms. Remote Sens. 2018, 10, 1513. [Google Scholar] [CrossRef] [Green Version]
  26. Franceschini, M.H.D.; Bartholomeus, H.; Van Apeldoorn, D.F.; Suomalainen, J.; Kooistra, L. Feasibility of unmanned aerial vehicle optical imagery for early detection and severity assessment of late blight in potato. Remote Sens. 2019, 11, 224. [Google Scholar] [CrossRef] [Green Version]
  27. Rodríguez, J.; Lizarazo, I.; Prieto, F.; Angulo–Morales, V. Assessment of potato late blight from UAV–based multispectral imagery. Comput. Electron. Agric. 2021, 184, 106061. [Google Scholar] [CrossRef]
  28. Aylor, D.E.; Fry, W.E.; Mayton, H.; Andrade–Piedra, J.L. Quantifying the rate of release and escape of Phytophthora infestans sporangia from a potato canopy. Phytopathology 2001, 91, 1189–1196. [Google Scholar] [CrossRef] [Green Version]
  29. Iglesias, I.; Escuredo, O.; Seijo, M.C.; Méndez, J. Phytophthora infestans prediction for a potato crop. Am. J. Potato Res. 2010, 87, 32–40. [Google Scholar] [CrossRef]
  30. Bugiani, R.; Govoni, P.; Bottazzi, R.; Giannico, P.; Montini, B.; Pozza, M. Monitoring airborne concentrations of sporangia of Phytophthora infestans in relation to tomato late blight in Emilia Romagna, Italy. Aerobiologia 1995, 11, 41–46. [Google Scholar] [CrossRef]
  31. Meno, L.; Abuley, I.K.; Escuredo, O.; Seijo, M.C. Suitability of Early Blight Forecasting Systems for Detecting First Symptoms in Potato Crops of NW Spain. Agronomy 2022, 12, 1611. [Google Scholar] [CrossRef]
  32. Escuredo, O.; Seijo, M.C.; Fernández–González, M.; Iglesias, I. Effects of meteorological factors on the levels of Alternaria spores on a potato crop. Int. J. Biometeorol. 2011, 55, 243–252. [Google Scholar] [CrossRef]
  33. Meno, L.; Escuredo, O.; Rodríguez–Flores, M.S.; Seijo, M.C. Modification of the tomcast model with aerobiological data for management of potato early blight. Agronomy 2020, 10, 1872. [Google Scholar] [CrossRef]
  34. Cao, K.; Ruckstuhl, M.; Forrer, H. Crucial Weather Conditions for Phytophthora infestans; PAGV-Special Report; Research Station for Arable Farming and Field Production of Vegetables: Lelystad, The Netherlands, 1997; Volume 85, p. 90. [Google Scholar]
  35. Skelsey, P.; Rossing, W.A.H.; Kessel, G.J.T.; Van der Werf, W. Scenario approach for assessing the utility of dispersal information in decision support for aerially spread plant pathogens, applied to Phytophthora infestans. Phytopathology 2009, 99, 887–895. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  36. Hack, H.; Gall, H.; Klemke, T.; Klose, R.; Meier, U.; Stauss, R.; Witzenberger, A. Phenological growth stages of potato (Solanum tuberosum L.). Nachrichtenbl. Deut. P£ Anzenschutzd 1993, 45, 11–19. [Google Scholar]
  37. Meno, L.; Escuredo, O.; Abuley, I.K.; Seijo, M.C. Importance of Meteorological Parameters and Airborne Conidia to Predict Risk of Alternaria on a Potato Crop Ambient Using Machine Learning Algorithms. Sensors 2022, 22, 7063. [Google Scholar] [CrossRef]
  38. Abuley, I.K.; Hansen, J.G.; Hansen, H.H. Controlling late blight in susceptible and resistant potato cultivars with BlightManager. In Applied Crop Protection; Danish Centre for Food and Agriculture; Aarhus Universitet: Aarhus, Denmark, 2020; Volume 187, pp. 92–97. Available online: https://dcapub.au.dk/djfpublikation/djfpdf/DCArapport187.pdf (accessed on 20 November 2021).
  39. Galán, S.C.; González, P.C.; Teno, P.A.; Vilches, E.D. Manual de Calidad y Gestión de la Red Española de Aerobiología; Universidad de Córdoba: Córdoba, Spain, 2007; 61p. [Google Scholar]
  40. R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2022. [Google Scholar]
  41. Nwanganga, F.; Chapple, M. Practical Machine Learning in R; John Wiley & Sons: Hoboken, NJ, USA, 2020; 464p. [Google Scholar]
  42. Kuhn, M.; Wing, J.; Weston, S.; Williams, A.; Keefer, C.; Engelhardt, A.; Cooper, T.; Mayer, Z.; Kenkel, B. Caret: Classification and Regression Training, R package version 6.0–86; Astrophysics Source Code Library: Cambridge, MA, USA, 2020. [Google Scholar]
  43. Sing, T.; Sander, O.; Beerenwinkel, N.; Lengauer, T. ROCR: Visualizing classifier performance in R. Bioinformatics 2005, 21, 3940–3941. [Google Scholar] [CrossRef] [PubMed]
  44. Iglesias, I.; Rodríguez–Rajo, F.J.; Méndez, J. Evaluation of the different Alternaria prediction models on a potato crop in A Limia (NW of Spain). Aerobiologia 2007, 23, 27–34. [Google Scholar] [CrossRef]
  45. Fenu, G.; Malloci, F.M. Artificial intelligence technique in crop disease forecasting: A case study on potato late blight prediction. In Proceedings of the International Conference on Intelligent Decision Technologies, Virtual Conference, 17–19 June 2020; pp. 79–89. [Google Scholar]
  46. Kaundal, R.; Kapoor, A.S.; Raghava, G.P. Machine learning techniques in disease forecasting: A case study on rice blast prediction. BMC Bioinform. 2006, 7, 485. [Google Scholar] [CrossRef] [Green Version]
  47. Hooker, D.; Schaafsma, A.; Tamburic–Ilincic, L. Using weather variables pre–and post–heading to predict deoxynivalenol content in winter wheat. Plant Dis. 2002, 86, 611–619. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  48. Klem, K.; Vanova, M.; Hajslova, J.; Lancová, K.; Sehnalová, M. A neural network model for prediction of deoxynivalenol content in wheat grain based on weather data and preceding crop. Plant Soil. Environ. 2007, 53, 421. [Google Scholar] [CrossRef] [Green Version]
  49. Fernández–González, M.; Rodríguez–Rajo, F.; Escuredo, O.; Aira, M. Optimization of integrated pest management for powdery mildew (Uncinula necator) control in a vineyard based on a combination of phenological, meteorological and aerobiological data. J. Agric. Sci. 2013, 151, 648–658. [Google Scholar] [CrossRef]
  50. Martínez–Bracero, M.; González–Fernández, E.; Wójcik, M.; Alcázar, P.; Fernández–González, M.; Kasprzyk, I.; Rodríguez–Rajo, F.; Galán, C. Airborne fungal phytopathological spore assessment in three European vineyards from different bioclimatic areas. Aerobiologia 2020, 36, 715–729. [Google Scholar] [CrossRef]
  51. Escuredo, O.; Seijo–Rodríguez, A.; Meno, L.; Rodríguez–Flores, M.S.; Seijo, M.C. Seasonal Dynamics of Alternaria during the Potato Growing Cycle and the Influence of Weather on the Early Blight Disease in North–West Spain. Am. J. Potato Res. 2019, 96, 532–540. [Google Scholar] [CrossRef]
  52. Meno; Escuredo, O.; Rodríguez–Flores, M.S.; Seijo, M.C. Looking for a sustainable potato crop. Field assessment of early blight management. Agric. For. Meteorol. 2021, 308–309, 108617. [Google Scholar]
  53. Rodríguez–Rajo, F.J.; Jato, V.; Fernandez–Gonzalez, M.; Aira, M.J. The use of aerobiological methods for forecasting Botrytis spore concentrations in a vineyard. Grana 2010, 49, 56–65. [Google Scholar] [CrossRef] [Green Version]
  54. Rotem, J.; Cohen, Y.; Putter, J. Relativity of limiting and optimum inoculum loads, wetting durations, and temperatures for infection by Phytophthora infestans. Phytopathology 1971, 61, 275–278. [Google Scholar] [CrossRef]
  55. Maziero, J.M.N.; Maffia, L.A.; Mizubuti, E.S. Effects of temperature on events in the infection cycle of two clonal lineages of Phytophthora infestans causing late blight on tomato and potato in Brazil. Plant Dis. 2009, 93, 459–466. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  56. Cooke, L.; Schepers, H.; Hermansen, A.; Bain, R.; Bradshaw, N.; Ritchie, F.; Shaw, D.; Evenhuis, A.; Kessel, G.; Wander, J. Epidemiology and integrated control of potato late blight in Europe. Potato Res. 2011, 54, 183–222. [Google Scholar] [CrossRef] [Green Version]
  57. Dar, W.; Parry, F.; Bhat, B. Potato late blight disease prediction using meteorological parameters in Northern Himalayas of India. J. Agrometeorol. 2021, 23, 310–315. [Google Scholar] [CrossRef]
Sensors 23 03818 g001 550
Figure 1. Visual abstract about the inputs (weather and aerobiological variables) used in the machine learning (ML) algorithms. ARL: daily aerobiological risk level.
Figure 1. Visual abstract about the inputs (weather and aerobiological variables) used in the machine learning (ML) algorithms. ARL: daily aerobiological risk level.
Sensors 23 03818 g001
Sensors 23 03818 g002 550
Figure 2. Receiver operating curve (ROC) for comparing the machine learning models random forest (RF) and C5.0 decision tree (C5.0). The broken lines in the ROC curve represent a classifier with no predictive value (diagonal line) and a perfect classifier (horizontal and vertical lines). ROC curves that are closer to the perfect classifier show a better model.
Figure 2. Receiver operating curve (ROC) for comparing the machine learning models random forest (RF) and C5.0 decision tree (C5.0). The broken lines in the ROC curve represent a classifier with no predictive value (diagonal line) and a perfect classifier (horizontal and vertical lines). ROC curves that are closer to the perfect classifier show a better model.
Sensors 23 03818 g002
Sensors 23 03818 g003 550
Figure 3. Comparison between prediction data by ML algorithms C5.0 and random forest (RF) and observed aerobiological risk level (ARL) by growing season. DAE: days after emergence.
Figure 3. Comparison between prediction data by ML algorithms C5.0 and random forest (RF) and observed aerobiological risk level (ARL) by growing season. DAE: days after emergence.
Sensors 23 03818 g003
Table
Table 1. Dates of phenological stages in five studied crop seasons.
Table 1. Dates of phenological stages in five studied crop seasons.
Date (DAE)
Stage20172018201920202021
Planting22 April15 May16 May26 May22 May
Emerging16 May (1)1 June (1)4 June (1)10 June (1)10 June (1)
Begining Flowering13 June (29)4 July (34)13 July (40)13 July (34)14 July (35)
Start of senescence19 July (65)15 August (76)17 August (75)14 August (66)7 August (59)
End of senescence12 August (89)14 September (106)1 September (90)31 August (83)26 August (78)
DAE: days after emergence.
Table
Table 2. Descriptive analysis of weather parameters by phenological stage during five studied crop seasons.
Table 2. Descriptive analysis of weather parameters by phenological stage during five studied crop seasons.
SeasonStageT (°C)RH (%)Accumulated Rain (mm)Rainy DaysRainy Days (%)
2017FD16.776.021.4631.6
FL19.672.9155.81052.6
SE17.769.684.2315.8
2018FD16.382.7123.42054.1
FL19.376.3121.4924.3
SE19.269.852.0821.6
2019FD15.776.942.81348.1
FL18.576.533.61037.0
SE18.972.44.0414.8
2020FD17.074.28.2228.6
FL21.465.44.2228.6
SE16.677.92.6342.9
2021FD16.380.768.01254.5
FL17.977.438.4836.4
SE19.273.22.429.1
Temperature (T) and relative humidity (RH) expressed in mean values.
Table
Table 3. Sporangia information by phenological stage during the five studied crop seasons.
Table 3. Sporangia information by phenological stage during the five studied crop seasons.
SeasonStageTotal P. infestans SporangiaDays ≥ 10 Sporangia/m3Days ≥ 10 Sporangia (%)
2017FD3111139.3
FL5211760.7
SE1500
2018FD17232555.6
FL5092044.4
SE3900
2019FD48111100
FL3700
SE500
2020FD2800
FL700
SE100
2021FD11342256.4
FL5401333.3
SE162410.3
FD: foliar development; FL: flowering; SE: senescence.
Table
Table 4. Spearman correlation coefficients between daily weather parameters and sporangia concentration until seven previous days.
Table 4. Spearman correlation coefficients between daily weather parameters and sporangia concentration until seven previous days.
Previous DaysTRHLWWindSREscapeIPSporangia
0−0.162 **0.198 **0.125 **0.020−0.179 **0.0300.258 **1
1−0.207 **0.218 **0.0810.026−0.169 **0.0620.259 **0.669 **
2−0.229 **0.221 **0.0390.158 **−0.156 **0.182 **0.256 **0.555 **
3−0.221 **0.243 **0.0720.153 **−0.180 **0.162 **0.247 **0.485 **
4−0.224 **0.249 **0.0690.146 **−0.185 **0.158 **0.254 **0.427 **
5−0.250 **0.268 **0.0780.165 **−0.198 **0.178 **0.252 **0.430 **
6−0.272 **0.322 **0.0890.106 *−0.247 **0.114 *0.247 **0.402 **
7−0.253 **0.259 **0.104 *0.056−0.200 **0.0660.224 **0.421 **
* p <0.05; ** p <0.01. T: mean temperature; RH: mean relative humidity; LW: leaf wetness; SR: spore release; IP: infection pressure.
Table
Table 5. Weather and sporangia variables in order of importance of selection according to ML algorithm.
Table 5. Weather and sporangia variables in order of importance of selection according to ML algorithm.
C5.0 AlgorithmRF Algorithm
Weather VariableUsage Rate (%)Weather VariableDecrease in Accuracy (%)
Wind_2100Sporangia_136
Wind_6100Sporangia_314
Sporangia_7100Sporangia_213
Sporangia_1100Sporangia_411
T_7100Sporangia_510
LW_0100Sporangia_710
IP_0100Sporangia_69
SR_5100SR_66
RH_7100Escape_26
RH_099IP_06
T: mean temperature; RH: mean relative humidity; LW: leaf wetness; SR: spore release; IP: infection pressure.
Table
Table 6. Relevant information about predictions by machine learning algorithms.
Table 6. Relevant information about predictions by machine learning algorithms.
AlgoritmAccuracyKappaSensitivitySpecificity
C5.00.850.650.810.87
RF0.870.700.770.92
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Meno, L.; Escuredo, O.; Abuley, I.K.; Seijo, M.C. Predicting Daily Aerobiological Risk Level of Potato Late Blight Using C5.0 and Random Forest Algorithms under Field Conditions. Sensors 2023, 23, 3818. https://doi.org/10.3390/s23083818

AMA Style

Meno L, Escuredo O, Abuley IK, Seijo MC. Predicting Daily Aerobiological Risk Level of Potato Late Blight Using C5.0 and Random Forest Algorithms under Field Conditions. Sensors. 2023; 23(8):3818. https://doi.org/10.3390/s23083818

Chicago/Turabian Style

Meno, Laura, Olga Escuredo, Isaac K. Abuley, and M. Carmen Seijo. 2023. "Predicting Daily Aerobiological Risk Level of Potato Late Blight Using C5.0 and Random Forest Algorithms under Field Conditions" Sensors 23, no. 8: 3818. https://doi.org/10.3390/s23083818

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop