Next Article in Journal
Population Development of Alien Gall-Forming Aphid Tetraneura nigriabdominalis on Ulmus minor and Ulmus glabra
Previous Article in Journal
Climatic Suitability and Distribution Overlap of Sawflies (Hymenoptera: Diprionidae) and Threatened Populations of Pinaceae
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Forests
Volume 13
Issue 7
10.3390/f13071068
forests-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

Machine Learning Methods to Estimate Productivity of Harvesters: Mechanized Timber Harvesting in Brazil

by
Rafaele Almeida Munis
1,
Rodrigo Oliveira Almeida
1,
Diego Aparecido Camargo
1,
Richardson Barbosa Gomes da Silva
1,
Jaime Wojciechowski
2 and
Danilo Simões
1,*
1
Department of Forest Science, Soils and Environment, School of Agriculture, São Paulo State University (UNESP), Botucatu 18610-034, Brazil
2
Informatics Department, Federal University of Paraná, Curitiba 81520-260, Brazil
*
Author to whom correspondence should be addressed.
Forests 2022, 13(7), 1068; https://doi.org/10.3390/f13071068
Submission received: 29 May 2022 / Revised: 2 July 2022 / Accepted: 4 July 2022 / Published: 7 July 2022
(This article belongs to the Section Forest Economics, Policy, and Social Science)
Browse Figures
Review Reports Versions Notes

Abstract

:
The correct capture of forest operations information carried out in forest plantations can help in the management of mechanized harvesting timber. Proper management must be able to dimension resources and tools necessary for the fulfillment of operations and helping in strategic, tactical, and operational planning. In order to facilitate the decision making of forest managers, this work aimed to analyze the performance of machine learning algorithms in estimating the productivity of timber harvesters. As predictors of productivity, we used the availability of hours of machine use, individual mean volumes of trees, and terrain slopes. The dataset was composed of 144,973 records, carried out over a period of 28 months. We tested the predictive performance of 24 machine learning algorithms in default mode. In addition, we tested the performance of blending and stacking joint learning methods. We evaluated the model’s fit using the root mean squared error, mean absolute error, mean absolute percentage error, and determination coefficient. After cleaning the initial database, we used only 1.12% to build the model. Learning by blending ensemble stood out with a determination coefficient of 0.71 and a mean absolute percentage error of 15%. From the use of data from machine learning algorithms, it became possible to predict the productivity of timber harvesters. Testing a variety of machine learning algorithms with different dynamics contributed to the machine learning technique that helped us reach our goal: maximizing the model’s performance by conducting experimentation.

1. Introduction

Management integrates the routine of forest managers responsible for guiding and implementing mechanized logging operations. The optimization of time and biological assets capitalized in planted forests, when exhausted by timber harvesters, affects the success of the operation. Thus, it is necessary to know the variables that influence mechanized timber harvesting, allowing for more effective planning.
Quality indicators, evaluation criteria, and risk analysis techniques enhance the structures that support decision makers. In doing so, data collected in forest inventory, measurements at the stand level, operational forest management, and onboard computers in forest machinery help and allow the management procedures for timber harvesting operations [1,2,3,4,5,6].
The manipulation and reuse of this information promote its use in the management development itself and helps in the identification of opportunities that can be foreseen. However, a mechanized timber harvesting operation planning scope necessarily requires a quantitative, robust, and reliable quantitative database [7,8].
On a daily basis, a large number of data are generated from the activities machines record that make up different harvesting systems. In Brazil, the mechanized cut-to-length (CTL) system, composed of a harvester and a forwarder, is normally used to supply raw material for forest-based industries. Due to the use of a harvester in the configuration of the CTL system, the trees are cut and sectioned into processed logs. Additionally, the timber without harvest residues, suitable for industrial processes, is arranged for extraction by means of a forwarder [9,10,11,12,13,14,15,16].
The sources and estimates reported by operators give the level of dataset quality. By providing early warnings, machine learning provides qualitative subsidies in detection of associations between variables, in addition to supporting hypotheses. However, machine learning algorithms must be associated with data wrangling methods to highlight outliers and reduce the size and facilitate dataset filtering [17,18,19,20,21].
In addition to data wrangling, the data preparation adapts to different analyses through the identification, extraction, cleansing, transformation, and integration of datasets. Data wrangling allows the user to explore data flow and problem distribution along the dataset to detect error patterns. This technique is conducive to the development of quality in the sequence of operations over time [22,23,24,25,26,27].
Forest managers, through datasets, are able to extract useful information and make better-informed decisions that include, for example, estimating the productivity of timber harvesters. Development of a model to predict dynamic productivity behavior of machines, under different configurations, can contribute to optimization of the operation [28,29,30,31].
Associating the knowledge generated from the application of data wrangling in a dataset with machine learning algorithms can promote intelligent decision making. The union of techniques of machine learning can improve the system performance of cut-to-length timber, forest resources, and services management [32,33,34,35].
Some studies investigated and proposed ways to improve this association, e.g., joint machine learning methods use a meta-model to combine predictive results from heterogeneous base models arranged in at least one layer. Thus, the variation in forecast combinations is based on a data validation set [36,37,38,39].
Learning methods through blending ensemble learning and stacking ensemble learning can promote more robust and accurate models because they decrease exponentially the error rate of joint learning. The blending ensemble learning method allows recognizing and training second-level algorithms to optimally combine prediction models, while stacking ensemble learning combines values by average or by additional models that add to the partial or final predictions of all individual models [40,41,42,43,44,45,46].
Machine learning models are increasingly used to predict attributes because, as they are not subject to traditional statistical assumptions, they can incorporate variables and find nonlinear and complex patterns in data. Machine learning includes a variety of algorithms for learning predictive rules from historical data and building models that can predict future values [47,48,49,50,51].
Machine learning has become an effective predictive analytics tool for large data volumes, with applications in medicine, finance, and, recently, the forestry sector. Forestry process automation has been proven to increase productivity and work quality. However, the mechanized timber harvesting operation is challenging and complex [52,53,54,55].
Traditionally, productivity is estimated through a time study, which consists of analyzing the work elements that make up the timber harvester’s operational cycle, individually, associating the time required and timber volume. However, the requirements of labor and minimum amount of sample needed make it expensive for forest-based industries.
As a result of these complex interactions, the use of efficient analytical tools by forest managers, such as machine learning, is justified to assist them in decision making. This work aimed to analyze the performance of machine learning algorithms in estimating the productivity of timber harvesters.

2. Materials and Methods

2.1. Dataset

We used structured data from the production and operation of mechanized timber harvesting in Eucalyptus- and Pinus-planted forests carried out by cut-to-length systems with harvesters. The planted forests with Eucalyptus had a spacing of 3.3 m × 1.8 m and mean age of 14 ± 9.87 years. The Pinus forest had a spacing of 3.3 m × 1.8 m and mean age of 22 ± 9.09 years. The wood from these forests was used as raw material for the production of pulp and paper.
The average meteorological conditions in the study region, according to the National Institute of Meteorology [56], were a relative humidity of 69.24%, wind speed of 4.29 ms−1, and an air temperature of 289.3 K. The operations took place in Brazil, in a region with a slope gradient from 7.32% to 35.06%. The intervals were categorized by a gentle (3% to 10%), moderate (10% to 32%), and steep slope relief (32% to 56%), according to Speight [57].
This research was based on empirical data and silvicultural inputs; therefore, the data were part of the daily records collected in the field by the onboard computers of timber harvesters. Despite considering all records in the initial analysis, we employed a series of compensatory controls from data wrangling.
In the two 10 h shifts daily, the machine availability, the individual mean volumes of trees, and the terrain slope added up to 144,973 instances incurred in the period of 28 months. These data were categorical, numerical, and ordinal, according to the box plot and distributions of predictor and target variables provided in the Supplementary Material (Figure S1). The bases were labeled, joined, and manipulated using the R programming language [58].
The actual times spent in activities were recorded in the onboard computers of timber harvesters. This way, we estimated productivity from the ratio between the timber volume extracted by a harvester, in cubic meters, and the effective operation time, in seconds [59,60]. The same operator could operate different brands and models of timber harvesters; however, this variable was not added in the construction of the model, due to the difficulty in tracking these data in the database. Altogether, the operating records of 21 harvesters were used (Table 1).
Through the programming language R, when implementing machine learning routines for management planning and detecting data quality, we considered, according to Konstantinou and Paton [61], procedures for transforming, cleaning, and merging different sources. We built a data wrangling routine, in which the outliers and potentially correlated variables were removed.
The instances went through the data wrangling process, which was performed in order to properly transform and gather acquired data. Additionally, through the interquartile range, conceptually defined by the Tukey range [62], we removed outliers and, using Spearman correlation, we verified the correlations between attributes (p < 0.05).
Furthermore, data balancing from SMOTE was performed. The SMOTE was adopted because it is a reference algorithm to solve the class disequilibrium learning problem [63]. The SMOTE algorithm has the dynamics of generating new synthetic examples in the neighborhood of small groups of nearby instances, using the k-nearest neighbor [64]. The function was implemented from the smotefamily package.

2.2. Different Learning Methods and Algorithm Approaches

Using a single dataset, we compared the predictive performance of 24 machine learning algorithms to estimate the productivity of timber harvesters. These algorithms were based on a decision tree, gradient boosting machine, linear regression, k-nearest neighbors, support vector machine, and artificial neural network.
For determining the best model, we used the metrics: root mean error (RMSE), mean absolute error (MAE), and mean absolute percent (MAPE). We used the determination coefficient (R2) as a final performance measure for each method.
We adopted the gradient (5, 10, 15, 20, and 25) for cross-validation, in which the hyperparameters were automatically optimized. Finally, we implemented stacking ensemble and blending ensemble learning methods. We ordered stacking ensemble learning methods in a hierarchical data structure. On each fold set, we applied k-fold cross-validation.
The predictive performance of models in relation to unseen data was maximized by determination coefficient (R2), minimizing test RMSE, which was determined from the random sample mean generated (n = 80). We implemented supervised learning regression using the Python programming language PyCaret library [65] to automate machine learning workflow and model development. From data instances, we grouped 90% (n = 1466) as a training set and 10% (n = 163) as a test set. We tested machine learning algorithms and selected them according to their performance in predicting productivity of forest machines, using universal statistical metrics for evaluating the performance of models [66], such as root mean squared error (RMSE), mean absolute error (MAE), mean absolute percent error (MAPE), and determination coefficient (R2).
Thus, we subjected the algorithms to different machine learning methods. First, we verified the performance of algorithms in a decoupled version, with hyperparameters in default mode. To improve performance, we adjusted the hyperparameters of the selected algorithms. We combined the validated data and formed the meta-feature set, the test data, and the target set. Again, we combined the sets using new meta-resource sets, creating a new meta-training set. The new target sets formed a new meta-test. We generated final predictions by meta-learner level one, from training with the meta-training set.
The combination learning method consisted of combining machine learning algorithms to minimize prediction error rates. For this, we divided the dataset into training and testing, as well as implementing zero-layer algorithms, which generated validation and test sets. We combined respective sets with new meta-training and meta-test sets [67] and generated final predictions by level-one meta-learner, from training with a meta-training set.

3. Results

3.1. Dataset Quality

The manipulation of the dataset with daily records of the mechanized timber harvesting operation resulted in a sample of 144,973 instances. However, because it was consolidated from the unity of different sources, including manual notes, the goodness of the dataset was partially compromised. Thus, we removed duplicate instances and instances with missing information.
With a data wrangling routine, in addition to cleaning filtering and transforming data, we carried out an examination of data quality, excluding outliers and promoting balancing. It is noteworthy that, despite timber harvesters having onboard computers, the data recording process still required manual interactions. Consequently, we implemented this process in 1.12% of the dataset, the models with machine learning algorithms (Table 2).
The attributes selected for the model building were individual mean volumes of trees, terrain slope, and availability of hours of machine use. More details about the mean, standard deviation, and median of the dataset from mechanized timber harvesting operation for attributes under study are shown in Table 3.

3.2. Different Learning Methods and Algorithm Approaches

First, we analyzed the predictive performance of 24 algorithms, individually, based on model fit metrics. Of the three trained algorithms, based on the decision tree, the determination coefficient of extra trees stood out, as it was 0.01 higher than the coefficient of determination of random forest and 0.22 higher than that of the decision tree (Table 4).
When analyzing the four algorithms based on gradient-boosted machines, the best determination coefficient was obtained by CatBoost Regressor, which was 0.04 higher than Gradient Boosting Regressor and 0.20 higher than AdaBoost Regressor (Table 5).
The algorithm’s availability based on linear regression contributed to the application of the twelve trained algorithms (Table 6).
It was found that the Automatic Relevance Determination, Kernel Ridge, Linear Regression, Huber Regression, Ridge Regression, and Bayesian Ridge algorithms showed the same determination coefficient, which was 0.02 higher than that of TheilSen Regressor, 0.06 higher than Least Angle Regression, 0.13 higher than Orthogonal Matching Pursuit, and 0.42 higher than Lasso Regression and Elastic Net.
Despite having different dynamics, the best determination coefficient was obtained by the k-neighbors regressor, which was 0.05 higher than the multi-layer perceptron regressor, 0.18 higher than the Random Sample Consensus, and 0.45 higher than Support Vector Regression (Table 7).
Among applied models that presented better determination coefficients were blending ensemble and stacking ensemble. Next, the algorithms were carried out in default mode, highlighting the Extra Trees Regressor (Table 8).
When analyzing the metrics in the dataset test, the blending ensemble model was confirmed as the best predictor of productivity of timber harvesters (Table 9).
In addition, as an assessment of overall model performance, we verified the 80 combinations of test set data, with the response. Thus, it was evident that the blending ensemble, followed by the stacking ensemble, produced relatively higher average values of R2 (Figure 1) and a lower degree of dispersion.
Figure 2 illustrates the performance of the main algorithms used in model construction to predict productivity, relating to observed values.
When we selected the black box algorithm, the increase in performance compromised the interpretation of the relationships between the predictor variables and the target variable. Complex mathematical functions made it difficult to infer from technical experts. However, by visualizing the distributions of predictor variables in each quartile of the response variable it was possible to infer that higher productivity was associated with greater machine availability and lower slope levels.
Although the algorithms used in the construction of models do not allow interpretability, as evidenced by productivity quartiles of the test set, the density distribution of predictor variables of individual mean volumes of trees, terrain slope, and machine availability was determined (Figure 3).

4. Discussion

Incorporating machine learning models into forest operations management routines allows managers to infer tactical and operational adjustments, with agility in decision making and accurate prognosis. In mechanized timber harvesting, the scenario dynamics and external influences that impact activities demand this adaptability together with forecasting capacities.
However, conducting and monitoring the performance of mechanized timber harvesting operations, using analytical tools such as machine learning, is restricted due to the quantity and quality of available data. Liski, et al. [31] and Maktoubian et al. [68]. Demirci et al. [69] and Abbasi et al. [70] report that decentralization and lack of data management in forest environments reduce the achievement of significant results.
The harvesters that acted as data sources had embedded technology, with output records of activities of timber cutting and sectioning. The lack of interoperability among electronic devices made communication and data transfer susceptible, compromising the possibility of instant corrections and perceptions of deviations in notes. Furthermore, Buccafurri et al. [71] and Shi et al. [72] point out that the quality of instances generated is part of a cooperative process, which requires participation, and therefore, leveling of all those involved. Data management must be aligned with operations’ organization, which makes it the responsibility of forest managers to coordinate these efforts.
The data residual volume, after execution of data wrangling processes, was still sufficient to verify the performance of machine learning algorithms in the productivity modeling of mechanized timber harvesting. Of the algorithm groups applied in the modeling, in the default process, the ones that performed best were those based on the decision tree, gradient-boosted machine, and k-nearest neighbor. Therefore, the best individual performance algorithms were, respectively: extract trees, gradient-boosted, and k-nearest neighbor.
There are many types of decision trees that have as their core the entropy of information. According to An and Zhou [73], in the specific analysis process, the gained information for each attribute is classified and ordered. Among the decision tree algorithms evaluated, the one that presented the best performance was the extremely randomized trees or extra trees algorithm. This algorithm was developed by Geurts et al., [74] and uses the same principle of random forests. However, as supported by Ahmad et al. [75], the extra trees may have differentiated themselves by using the entire training dataset to train each regression tree and not just a bootstrap replica.
Of the algorithms based on gradient-boosted machines, the CatBoost Regressor showed the best fit model to the data. This algorithm developed by Prokhorenkova et al. [76] is an enhancement of gradient boosting, designed to avoid attribute dependency and improve prediction accuracy on small datasets. As it is a non-parametric algorithm that, according to Ortiz-Bejar et al. [77], stores all known observations and uses them in the prediction based on similarity functions, the third-best performance was from the model based on k-nearest neighbor.
As a way of enhancing the prediction, tests were carried out with the blending ensemble and stacking ensemble learning methods, using combined learning. These learnings were combined from the three algorithms, in default mode, which presented the best performances. The predictions obtained by both methods were superior to those obtained by algorithms in default mode. Jong et al. [78] and Jordan and Mitchell [79] point out that, in general, combined learning methods increase the performance of models built with machine learning.
Associating the blending ensemble use with the possibility of pre-determining productivity, based on attributes of individual mean volumes of trees, terrain slope, and availability of hours of machine use, promotes dynamism in managers’ planning, especially in operational planning, which requires quick responses in adverse operating conditions. This corroborates the limitations of traditional estimating method productivity through the study of times.
In addition, the comparison through values of employed models’ scatter diagrams demonstrated the effects of predictor variables on productivity. In upper quartiles, in operating conditions with lower slopes and longer availability harvesters, their effects increase considerably the productivity.
The building of models involving machine learning algorithms, in addition to providing prediction of harvester productivity in the mechanized timber harvesting operation, allowed us to look at the bases that guide strategic decisions of operations in planted forests. This opportunity has shown that, despite the quality, suitable data promote knowledge extraction, mainly from attributes not correlated with productivity.

5. Conclusions

From the use of adjusted data of machine learning algorithms, it is possible to predict the productivity of timber harvesters.
Among the attributes that compose datasets of mechanized timber harvesting activities, the individual mean volumes of trees, terrain slope, and machine availability are the main factors that impact harvester productivity estimation.
Testing a variety of machine learning algorithms with different dynamics contributed to the development of a machine learning technique that enabled what it proposes, i.e., experimentation and good performance of the models. Thus, the choice for blending ensemble learning was guided by the comparison of model fit statistical metrics.
Among the learning methods by blending ensemble, stacking ensemble, and algorithms, in default mode, the blending ensemble had a determination coefficient of 0.71 and a mean absolute percent error of 15%.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/f13071068/s1. Figure S1: Raincloud plot with box plot and the distributions of predictor variables and target (A) individual mean volumes of trees; (B) terrain slope; (C) machine availability; and (D) productivity.

Author Contributions

Conceptualization, R.A.M., D.S. and J.W.; methodology, R.A.M., D.S., R.O.A. and D.A.C.; software, R.A.M. and R.O.A.; validation, R.A.M., R.O.A., D.A.C., R.B.G.d.S., J.W. and D.S.; formal analysis, R.A.M., R.O.A. and D.A.C.; investigation, R.A.M., D.A.C. and R.B.G.d.S.; data curation, R.A.M. and R.O.A.; writing—original draft preparation, R.A.M., D.A.C. and R.B.G.d.S.; writing—review and editing, R.A.M., R.O.A., D.A.C., R.B.G.d.S., J.W. and D.S.; visualization, supervision, D.S. and J.W.; project administration, D.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

All data are provided in the main manuscript. Contact the corresponding author if further explanation is required.

Acknowledgments

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq).

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Pollard, S.J.T.; Brookes, A.; Earl, N.; Lowe, J.; Kearney, T.; Nathanail, C.P. Integrating decision tools for the sustainable management of land contamination. Sci. Total Environ. 2004, 325, 15–28. [Google Scholar] [CrossRef] [Green Version]
  2. Bai, C.; Sarkis, J. Integrating and extending data and decision tools for sustainable third-party reverse logistics provider selection. Comput. Oper. Res. 2019, 110, 188–207. [Google Scholar] [CrossRef]
  3. Welch, H.; Brodie, S.; Jacox, M.G.; Bograd, S.J.; Hazen, E.L. Decision-support tools for dynamic management. Conserv. Biol. 2020, 34, 589–599. [Google Scholar] [CrossRef]
  4. McRoberts, R.E.; Tomppo, E.O. Remote sensing support for national forest inventories. Remote Sens. Environ. 2007, 110, 412–419. [Google Scholar] [CrossRef]
  5. Liang, X.; Kankare, V.; Hyyppä, J.; Wang, Y.; Kukko, A.; Haggrén, H.; Yu, X.; Kaartinen, H.; Jaakkola, A.; Guan, F.; et al. Terrestrial laser scanning in forest inventories. ISPRS J. Photogramm. Remote Sens. 2016, 115, 63–77. [Google Scholar] [CrossRef]
  6. Eyvindson, K.; Saad, R.; Eriksson, L.O. Incorporating stand level risk management options into forest decision support systems. For. Syst. 2017, 26, e013. [Google Scholar] [CrossRef]
  7. Wagner, J.E. Misinterpreting the internal rate of return in sustainable forest management planning and economic analysis. J. Sustain. For. 2012, 31, 239–266. [Google Scholar] [CrossRef]
  8. Karttunen, K.; Laitila, J. Forest management regime options for integrated small-diameter wood harvesting and supply chain from young Scots pine (Pinus sylvestris L.) stands. Int. J. For. Eng. 2015, 26, 124–138. [Google Scholar] [CrossRef]
  9. Camargo, D.A.; Munis, R.A.; Simões, D. Investigation of exposure to occupational noise among forestry machine operators: A case study in Brazil. Forests 2021, 12, 299. [Google Scholar] [CrossRef]
  10. Fernandez-Lacruz, R.; Edlund, M.; Bergström, D.; Lindroos, O. Productivity and profitability of harvesting overgrown roadside verges—A Swedish case study. Int. J. For. Eng. 2021, 32, 19–28. [Google Scholar] [CrossRef]
  11. Visser, R.; Berkett, H. Effect of terrain steepness on machine slope when harvesting. Int. J. For. Eng. 2015, 26, 1–9. [Google Scholar] [CrossRef]
  12. Sherwin, L.M.; Owende, P.M.O.; Kanali, C.L.; Lyons, J.; Ward, S.M. Influence of tyre inflation pressure on whole-body vibrations transmitted to the operator in a cut-to-length timber harvester. Appl. Ergon. 2004, 35, 253–261. [Google Scholar] [CrossRef]
  13. Ovaskainen, H.; Heikkilä, M. Visuospatial cognitive abilities in cut-to-length single-grip timber harvester work. Int. J. Ind. Ergon. 2007, 37, 771–780. [Google Scholar] [CrossRef]
  14. Walsh, D.; Strandgard, M. ScienceDirect Productivity and cost of harvesting a stemwood biomass product from integrated cut-to-length harvest operations in Australian Pinus radiata plantations. Biomass Bioenergy 2014, 738, 93–102. [Google Scholar] [CrossRef]
  15. Hera, P.L.; Morales, D.O.; Mendoza-Trejo, O. A study case of Dynamic Motion Primitives as a motion planning method to automate the work of forestry cranes. Comput. Electron. Agric. 2021, 183, 106037. [Google Scholar] [CrossRef]
  16. Huang, X. Application analysis of AI reasoning engine in microblog culture industry. Pers. Ubiquitous Comput. 2020, 24, 393–403. [Google Scholar] [CrossRef]
  17. Cho, G.; Park, H.-M.; Jung, W.-M.; Cha, W.-S.; Lee, D.; Chae, Y. Identification of candidate medicinal herbs for skincare via data mining of the classic Donguibogam text on Korean medicine. Integr. Med. Res. 2020, 9, 100436. [Google Scholar] [CrossRef] [PubMed]
  18. Rodrigues de Holanda Maia, M.; Plastino, A.; Penna, P.H.V. MineReduce: An approach based on data mining for problem size reduction. Comput. Oper. Res. 2020, 122, 104995. [Google Scholar] [CrossRef]
  19. Xu, Z.; Cheng, X.; Wang, K.; Yang, S. Analysis of the environmental trend of network finance and its influence on traditional commercial banks. J. Comput. Appl. Math. 2020, 379, 112907. [Google Scholar] [CrossRef]
  20. da Silva, A.K.V.; Borges, M.V.V.; Batista, T.S.; da Junior, C.A.S.; Furuya, D.E.G.; Osco, L.P.; Teodoro, L.P.R.; Baio, F.H.R.; Ramos, A.P.M.; Gonçalves, W.N.; et al. Predicting eucalyptus diameter at breast height and total height with uav-based spectral indices and machine learning. Forests 2021, 12, 582. [Google Scholar] [CrossRef]
  21. Giannetti, F.; Pecchi, M.; Travaglini, D.; Francini, S.; D’amico, G.; Vangi, E.; Cocozza, C.; Chirici, G. Estimating vaia windstorm damaged forest area in italy using time series sentinel-2 imagery and continuous change detection algorithms. Forests 2021, 12, 680. [Google Scholar] [CrossRef]
  22. Kandel, S.; Heer, J.; Plaisant, C.; Kennedy, J.; Van Ham, F.; Riche, N.H.; Weaver, C.; Lee, B.; Brodbeck, D.; Buono, P. Research directions in data wrangling: Visualizations and transformations for usable and credible data. Inf. Vis. 2011, 10, 271–288. [Google Scholar] [CrossRef] [Green Version]
  23. Endel, F.; Piringer, H. Data Wrangling: Making data useful again. IFAC-PapersOnLine 2015, 28, 111–112. [Google Scholar] [CrossRef]
  24. Furche, T.; Gottlob, G.; Libkin, L.; Orsi, G.; Paton, N.W. Data wrangling for big data: Challenges and opportunities. Adv. Database Technol.-EDBT 2016, 2016, 473–478. [Google Scholar] [CrossRef]
  25. Bellomarini, L.; Fayzrakhmanov, R.R.; Gottlob, G.; Kravchenko, A.; Laurenza, E.; Nenov, Y.; Reissfelder, S.; Sallinger, E.; Sherkhonov, E.; Vahdati, S.; et al. Data science with Vadalog: Knowledge Graphs with machine learning and reasoning in practice. Futur. Gener. Comput. Syst. 2022, 129, 407–422. [Google Scholar] [CrossRef]
  26. Bors, C.; Gschwandtner, T.; Miksch, S. Capturing and visualizing provenance from data wrangling. IEEE Comput. Graph. Appl. 2019, 39, 61–75. [Google Scholar] [CrossRef]
  27. Yang, J.; Li, R.; Chen, L.; Hu, Y.; Dou, Z. Research on equipment corrosion diagnosis method and prediction model driven by data. Process Saf. Environ. Prot. 2022, 158, 418–431. [Google Scholar] [CrossRef]
  28. De Jaegher, B.; Larumbe, E.; De Schepper, W.; Verliefde, A.; Nopens, I. Colloidal fouling in electrodialysis: A neural differential equations model. Sep. Purif. Technol. 2020, 249, 116939. [Google Scholar] [CrossRef]
  29. Kudyba, S. A hybrid analytic approach for understanding patient demand for mental health services. Netw. Model. Anal. Health Inform. Bioinforma. 2018, 7, 3. [Google Scholar] [CrossRef]
  30. Liu, J.; Kadziński, M.; Liao, X.; Mao, X.; Wang, Y. A preference learning framework for multiple criteria sorting with diverse additive value models and valued assignment examples. Eur. J. Oper. Res. 2020, 286, 963–985. [Google Scholar] [CrossRef] [Green Version]
  31. Liski, E.; Jounela, P.; Korpunen, H.; Sosa, A.; Lindroos, O.; Jylhä, P. Modeling the productivity of mechanized CTL harvesting with statistical machine learning methods. Int. J. For. Eng. 2020, 31, 253–262. [Google Scholar] [CrossRef]
  32. Sunhare, P.; Chowdhary, R.R.; Chattopadhyay, M.K. Internet of things and data mining: An application oriented survey. J. King Saud Univ.-Comput. Inf. Sci. 2020, 34, 3569–3590. [Google Scholar] [CrossRef]
  33. Luo, M.; Wang, Y.; Xie, Y.; Zhou, L.; Qiao, J.; Qiu, S.; Sun, Y. Combination of feature selection and catboost for prediction: The first application to the estimation of aboveground biomass. Forests 2021, 12, 216. [Google Scholar] [CrossRef]
  34. Koreň, M.; Jakuš, R.; Zápotocký, M.; Barka, I.; Holuša, J.; Ďuračiová, R.; Blaženec, M. Assessment of machine learning algorithms for modeling the spatial distribution of bark beetle infestation. Forests 2021, 12, 395. [Google Scholar] [CrossRef]
  35. Kusiak, A. Data mining: Manufacturing and service applications. Int. J. Prod. Res. 2006, 44, 4175–4191. [Google Scholar] [CrossRef]
  36. Maxwell, A.E.; Warner, T.A.; Strager, M.P.; Conley, J.F.; Sharp, A.L. Assessing machine-learning algorithms and image- and lidar-derived variables for GEOBIA classification of mining and mine reclamation. Int. J. Remote Sens. 2015, 36, 954–978. [Google Scholar] [CrossRef]
  37. Taylor, P.; Griffiths, N.; Bhalerao, A.; Anand, S.; Popham, T.; Xu, Z.; Gelencser, A. Data Mining for Vehicle Telemetry. Appl. Artif. Intell. 2016, 30, 233–256. [Google Scholar] [CrossRef] [Green Version]
  38. Chatzimparmpas, A.; Martins, R.M.; Kucher, K.; Kerren, A. Empirical Study: Visual Analytics for Comparing Stacking to Blending Ensemble Learning. In Proceedings of the 2021 23rd International Conference on Control Systems and Computer Science (CSCS), Bucharest, Romania, 26–28 May 2021; pp. 1–8. [Google Scholar] [CrossRef]
  39. Hansrajh, A.; Adeliyi, T.T.; Wing, J. Detection of Online Fake News Using Blending Ensemble Learning. Sci. Program. 2021, 2021, 3434458. [Google Scholar] [CrossRef]
  40. Wu, T.; Zhang, W.; Jiao, X.; Guo, W.; Alhaj Hamoud, Y. Evaluation of stacking and blending ensemble learning methods for estimating daily reference evapotranspiration. Comput. Electron. Agric. 2021, 184, 106039. [Google Scholar] [CrossRef]
  41. Divina, F.; Gilson, A.; Goméz-Vela, F.; Torres, M.G.; Torres, J.F. Stacking ensemble learning for short-term electricity consumption forecasting. Energies 2018, 11, 949. [Google Scholar] [CrossRef] [Green Version]
  42. Malik, F.A.; Ye, W.; Chen, Q.; Li, D. Recommendation algorithm based on blending learning. In Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference, Guangzhou, China, 22–24 June 2019; pp. 113–117. [Google Scholar] [CrossRef]
  43. Fang, Z.; Wang, Y.; Peng, L.; Hong, H. A comparative study of heterogeneous ensemble-learning techniques for landslide susceptibility mapping. Int. J. Geogr. Inf. Sci. 2021, 35, 321–347. [Google Scholar] [CrossRef]
  44. Hao, M.; Cao, W.H.; Liu, Z.T.; Wu, M.; Xiao, P. Visual-audio emotion recognition based on multi-task and ensemble learning with multiple features. Neurocomputing 2020, 391, 42–51. [Google Scholar] [CrossRef]
  45. Sun, W.; Trevor, B. A stacking ensemble learning framework for annual river ice breakup dates. J. Hydrol. 2018, 561, 636–650. [Google Scholar] [CrossRef]
  46. Lee, J.; Kim, J.; Ko, W. Day-ahead electric load forecasting for the residential building with a small-size dataset based on a self-organizing map and a stacking ensemble learning method. Appl. Sci. 2019, 9, 1231. [Google Scholar] [CrossRef] [Green Version]
  47. Ünver-Okan, S. Modelling of work efficiency in cable traction with tractor implementing the least-squares methods and robust regression. Croat. J. For. Eng. 2020, 41, 109–117. [Google Scholar] [CrossRef]
  48. Arumugam, K.; Swathi, Y.; Sanchez, D.T.; Mustafa, M.; Phoemchalard, C.; Phasinam, K.; Okoronkwo, E. Towards applicability of machine learning techniques in agriculture and energy sector. Mater. Today Proc. 2022, 51, 2260–2263. [Google Scholar] [CrossRef]
  49. Morera, A.; Martínez de Aragón, J.; De Cáceres, M.; Bonet, J.A.; De-Miguel, S. Historical and future spatially-explicit climate change impacts on mycorrhizal and saprotrophic macrofungal productivity in Mediterranean pine forests. Agric. For. Meteorol. 2022, 319, 108918. [Google Scholar] [CrossRef]
  50. Jin, Z.; Shang, J.; Zhu, Q.; Ling, C.; Xie, W.; Qiang, B. RFRSF: Employee Turnover Prediction Based on Random Forests and Survival Analysis. In Web Information Systems Engineering—WISE 2020; Lecture Notes in Computer Science; Huang, Z., Beek, W., Wang, H., Zhou, R., Zhang, Y., Eds.; Springer: Cham, Switzerland, 2020; Volume 123, pp. 503–515. [Google Scholar] [CrossRef]
  51. Chen, Y.; Dong, C.; Wu, B. Crown Profile Modeling and Prediction Based on Ensemble Learning. Forests 2022, 13, 410. [Google Scholar] [CrossRef]
  52. Palonen, T.; Hyyti, H.; Visala, A. Augmented Reality in Forest Machine Cabin. IFAC-PapersOnLine 2017, 50, 5410–5417. [Google Scholar] [CrossRef]
  53. Marčeta, D.; Petković, V.; Ljubojević, D.; Potočnik, I. Harvesting System Suitability as Decision Support in Selection Cutting Forest Management in Northwest Bosnia and Herzegovina. Croat. J. For. Eng. 2020, 41, 251–265. [Google Scholar] [CrossRef]
  54. Aworka, R.; Cedric, L.S.; Adoni, W.Y.H.; Zoueu, J.T.; Mutombo, F.K.; Kimpolo, C.L.M.; Nahhal, T.; Krichen, M. Agricultural decision system based on advanced machine learning models for yield prediction: Case of East African countries. Smart Agric. Technol. J 2022, 2, 100048. [Google Scholar] [CrossRef]
  55. Kamarulzaman, A.M.M.; Jaafar, W.S.W.M.; Maulud, K.N.A.; Saad, S.N.M.; Omar, H.; Mohan, M. Integrated Segmentation Approach with Machine Learning Classifier in Detecting and Mapping Post Selective Logging Impacts Using UAV Imagery. Forests 2022, 13, 48. [Google Scholar] [CrossRef]
  56. National Institute of Meteorology. Available online: https//portal.inmet.gov.br/dadoshistoricos (accessed on 18 February 2021).
  57. Speight, J.G.; Isbell, R.F. Soil Profiles. In Australian Soil and Land Survey Field Handbook; CSIRO: Canberra, Australia, 2009; pp. 132–133. ISBN 9780643093959. [Google Scholar]
  58. R Development Core Team. A Language and Environment for Statistical Computing; R Foundation for Computing Statistical: Vienna, Austria, 2021; ISBN 3-900051-07-0. [Google Scholar]
  59. Suzuki, Y.; Yoshimura, T. Assessment of broad-leaved forest stand management: Stock densities, thinning costs and profits over a 60-year rotation period. Croat. J. For. Eng. 2019, 40, 365–375. [Google Scholar] [CrossRef]
  60. Lemm, R.; Blattert, C.; Holm, S.; Bont, L.; Thees, O. Improving economic management decisions in forestry with the sorsim assortment model. Croat. J. For. Eng. 2020, 41, 71–83. [Google Scholar] [CrossRef] [Green Version]
  61. Konstantinou, N.; Paton, N.W. Feedback driven improvement of data preparation pipelines. Inf. Syst. 2020, 92, 101480. [Google Scholar] [CrossRef] [Green Version]
  62. Tukey, J.W. Exploratory Data Analysis, 1st ed.; Pearson: London, UK, 1977; ISBN 978-0201076165. [Google Scholar]
  63. Zhang, A.; Yu, H.; Zhou, S.; Huan, Z.; Yang, X. Instance weighted SMOTE by indirectly exploring the data distribution. Knowl.-Based Syst. 2022, 249, 108919. [Google Scholar] [CrossRef]
  64. Juez-Gil, M.; Arnaiz-González, Á.; Rodríguez, J.J.; López-Nozal, C.; García-Osorio, C. Approx-SMOTE: Fast SMOTE for Big Data on Apache Spark. Neurocomputing 2021, 464, 432–437. [Google Scholar] [CrossRef]
  65. PyCaret Org. 2021. Available online: https://pycaret.org (accessed on 18 February 2021).
  66. Mitchell, T.M. Machine Learning, 1st ed.; McGraw-Hill Science/Engineering/Math: New York, NY, USA, 1977; ISBN 978-0070428072. [Google Scholar]
  67. Lin, X.; Wu, J.; Wei, Y. An ensemble learning velocity prediction-based energy management strategy for a plug-in hybrid electric vehicle considering driving pattern adaptive reference SOC. Energy 2021, 234, 121308. [Google Scholar] [CrossRef]
  68. Maktoubian, J.; Taskhiri, M.S.; Turner, P. Intelligent predictive maintenance (Ipdm) in forestry: A review of challenges and opportunities. Forests 2021, 12, 1495. [Google Scholar] [CrossRef]
  69. Demirci, M.; Yesil, A.; Bettinger, P. Introducing a New Approach in Stand Tending Planning and Thinning Block Designation by Using Mixed Integer Goal Programming. Croat. J. For. Eng. 2022, 43, 134–151. [Google Scholar] [CrossRef]
  70. Abbasi, R.; Martinez, P.; Ahmad, R. The digitization of agricultural industry—A systematic literature review on agriculture 4.0. Smart Agric. Technol. 2022, 2, 100042. [Google Scholar] [CrossRef]
  71. Buccafurri, F.; De Meo, P.; Fugini, M.; Furnari, R.; Goy, A.; Lax, G.; Lops, P.; Modafferi, S.; Pernici, B.; Redavid, D.; et al. Analysis of QoS in cooperative services for real time applications. Data Knowl. Eng. 2008, 67, 463–484. [Google Scholar] [CrossRef]
  72. Shi, M.; Xu, J.; Liu, S.; Xu, Z. Productivity-Based Land Suitability and Management Sensitivity Analysis: The Eucalyptus E. urophylla × E. grandis Case. Forests 2022, 13, 340. [Google Scholar] [CrossRef]
  73. An, Y.; Zhou, H. Short term effect evaluation model of rural energy construction revitalization based on ID3 decision tree algorithm. Energy Rep. 2022, 8, 1004–1012. [Google Scholar] [CrossRef]
  74. Geurts, P.; Ernst, D.; Wehenkel, L. Extremely randomized trees. Mach. Learn. 2006, 63, 3–42. [Google Scholar] [CrossRef] [Green Version]
  75. Ahmad, M.W.; Reynolds, J.; Rezgui, Y. Predictive modelling for solar thermal energy systems: A comparison of support vector regression, random forest, extra trees and regression trees. J. Clean. Prod. 2018, 203, 810–821. [Google Scholar] [CrossRef]
  76. Prokhorenkova, L.; Gusev, G.; Vorobev, A.; Dorogush, A.V.; Gulin, A. Catboost: Unbiased boosting with categorical features. Adv. Neural Inf. Process. Syst. 2018, 2018, 6638–6648. [Google Scholar]
  77. Ortiz-Bejar, J.; Graff, M.; Tellez, E.S.; Ortiz-Bejar, J.; Jacobo, J.C. κ-Nearest neighbor regressors optimized by using random search. In Proceedings of the 2018 IEEE International Autumn Meeting on Power, Electronics and Computing (ROPEC), Ixtapa, Mexico, 14–16 November 2018. [Google Scholar] [CrossRef]
  78. Jong, N.; Krumeich, J.S.M.; Verstegen, D.M.L. To what extent can PBL principles be applied in blended learning: Lessons learned from health master programs. Med. Teach. 2017, 39, 203–211. [Google Scholar] [CrossRef] [Green Version]
  79. Jordan, M.I.; Mitchell, T.M. Machine learning: Trends, perspectives, and prospects. Science 2015, 349, 255–260. [Google Scholar] [CrossRef]
Forests 13 01068 g001 550
Figure 1. Average determination coefficient values of employed models.
Figure 1. Average determination coefficient values of employed models.
Forests 13 01068 g001
Forests 13 01068 g002 550
Figure 2. Comparison of scatter plot values of employed models.
Figure 2. Comparison of scatter plot values of employed models.
Forests 13 01068 g002
Forests 13 01068 g003 550
Figure 3. Density distribution of employed models.
Figure 3. Density distribution of employed models.
Forests 13 01068 g003
Table
Table 1. Technical specifications of timber harvesters used in cutting and processing trees.
Table 1. Technical specifications of timber harvesters used in cutting and processing trees.
Timber HarvestersQuantities of Timber Harvesters
BrandModelPower (kW)
Harvester PonsseERGO 8W20517
Harvester John Deere1270E1702
Harvester PonsseBEAVER1501
Harvester John Deere1270G2001
Table
Table 2. Data wrangling procedures applied to the dataset from the mechanized timber harvesting operation.
Table 2. Data wrangling procedures applied to the dataset from the mechanized timber harvesting operation.
Data Wrangling ProceduresInputOutput
Duplicate instances144,973144,781
Removal of missing data144,78198,459
Annotation errors98,4591703
Outlier removal17031629
Table
Table 3. Mean, standard deviation, and median of dataset from mechanized timber harvesting operation, after process of removing outliers of three initial attributes.
Table 3. Mean, standard deviation, and median of dataset from mechanized timber harvesting operation, after process of removing outliers of three initial attributes.
AttributeMinimumMaximumMeanStandard
Deviation		MedianSkewnessKurtosis
Individual Mean Volumes of Trees [m3]0.110.620.350.090.360.092.83
Terrain Slope [%]2.8635.0612.376.4611.401.626.26
Availability of Hours of Machine use [h]0.151.330.690.120.70−0.555.19
Table
Table 4. Evaluation metric of models based on decision tree applied to the training set from mechanized timber harvesting operation.
Table 4. Evaluation metric of models based on decision tree applied to the training set from mechanized timber harvesting operation.
ModelMAE
[m3 h−1]		RMSE
[m3 h−1]		R2MAPE
[%]		Time
[s]
Extra Trees Regressor3.214.680.72170.12
Random Forest Regressor3.364.720.71180.13
Decision Tree Regressor4.266.220.50220.01
MAE: mean absolute error; RMSE: root mean squared error; R2: determination coefficient; MAPE: mean absolute percent error.
Table
Table 5. Evaluation metric of models based on gradient-boosted machine applied to the training set from mechanized timber harvesting operation.
Table 5. Evaluation metric of models based on gradient-boosted machine applied to the training set from mechanized timber harvesting operation.
ModelMAE
[m3 h−1]		RMSE
[m3 h−1]		R2MAPE
[%]		Time
[s]
CatBoost Regressor3.354.550.73181.40
Gradient Boosting Regressor3.684.910.69200.04
AdaBoost Regressor4.826.040.53280.03
MAE: mean absolute error; RMSE: root mean squared error; R2: determination coefficient; MAPE: mean absolute percent error.
Table
Table 6. Evaluation metric of models based on linear regression applied to the training set from mechanized timber harvesting operation.
Table 6. Evaluation metric of models based on linear regression applied to the training set from mechanized timber harvesting operation.
ModelMAE
[m3 h−1]		RMSE
[m3 h−1]		R2MAPE
[%]		Time
[s]
Automatic Relevance Determination4.626.020.53250.03
Kernel Ridge4.656.040.53250.02
Linear Regression4.626.030.53250.30
Huber Regressor4.586.050.53250.02
Ridge Regression4.646.040.53250.20
Bayesian Ridge4.626.030.53250.02
TheilSen Regressor4.726.190.51260.51
Least Angle Regression4.876.380.47260.01
Orthogonal Matching Pursuit5.496.870.40290.01
Lasso Regression6.688.350.11400.28
Elastic Net6.688.340.11400.02
Lasso Least Angle Regression7.328.860.00440.02
MAE: mean absolute error; RMSE: root mean squared error; R2: determination coefficient; MAPE: mean absolute percent error.
Table
Table 7. Evaluation metric of models based on k-nearest neighbor, multi-layer perception regressor, random sample consensus, support vector regression, dummy regressor, and passive-aggressive regressor applied to the training set from mechanized timber harvesting operation.
Table 7. Evaluation metric of models based on k-nearest neighbor, multi-layer perception regressor, random sample consensus, support vector regression, dummy regressor, and passive-aggressive regressor applied to the training set from mechanized timber harvesting operation.
ModelMAE
[m3 h−1]		RMSE
[m3 h−1]		R2MAPE
[%]		Time
[s]
K-Neighbors Regressor4.095.590.60240.02
Multilayer Perceptron Regressor4.575.940.55250.40
Random Sample Consensus5.106.730.42270.02
Support Vector Regression6.358.180.15370.03
Dummy Regressor7.328.860.00440.01
Passive-Aggressive Regressor8.2510.02−0.49520.02
MAE: mean absolute error; RMSE: root mean squared error; R2: determination coefficient; MAPE: mean absolute percent error.
Table
Table 8. Performance obtained on the training set, by default, stacking ensemble learning and blending ensemble learning model.
Table 8. Performance obtained on the training set, by default, stacking ensemble learning and blending ensemble learning model.
ModelProcessMAE
[m3 h−1]		RMSE
[m3 h−1]		R2MAPE
[%]
Extra Trees Regressor (A)Default3.314.490.7418
CatBoost Regressor (B)Default3.134.580.7317
K-Neighbors Regressor (C)Default4.065.540.6124
A, B e CBlending ensemble3.184.430.7518
A, B e C–AStacking ensemble3.214.430.7518
MAE: mean absolute error; RMSE: root mean squared error; R2: determination coefficient; MAPE: mean absolute percent error.
Table
Table 9. Results obtained on the test dataset, by default, stacking ensemble learning and blending ensemble learning model.
Table 9. Results obtained on the test dataset, by default, stacking ensemble learning and blending ensemble learning model.
ModelProcessR2MAPE [%]
A, B e CBlending ensemble0.7715
A, B e C–AStacking ensemble0.7614
Extra Trees RegressorDefault0.7414
CatBoost RegressorDefault0.7415
K-Neighbors RegressorDefault0.6719
R2: determination coefficient; MAPE: mean absolute percent error.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Munis, R.A.; Almeida, R.O.; Camargo, D.A.; da Silva, R.B.G.; Wojciechowski, J.; Simões, D. Machine Learning Methods to Estimate Productivity of Harvesters: Mechanized Timber Harvesting in Brazil. Forests 2022, 13, 1068. https://doi.org/10.3390/f13071068

AMA Style

Munis RA, Almeida RO, Camargo DA, da Silva RBG, Wojciechowski J, Simões D. Machine Learning Methods to Estimate Productivity of Harvesters: Mechanized Timber Harvesting in Brazil. Forests. 2022; 13(7):1068. https://doi.org/10.3390/f13071068

Chicago/Turabian Style

Munis, Rafaele Almeida, Rodrigo Oliveira Almeida, Diego Aparecido Camargo, Richardson Barbosa Gomes da Silva, Jaime Wojciechowski, and Danilo Simões. 2022. "Machine Learning Methods to Estimate Productivity of Harvesters: Mechanized Timber Harvesting in Brazil" Forests 13, no. 7: 1068. https://doi.org/10.3390/f13071068

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