Current issue: 58(4)

Scopus CiteScore 2023: 3.5
Scopus ranking of open access forestry journals: 17th
PlanS compliant
Select issue
Silva Fennica 1926-1997
1990-1997
1980-1989
1970-1979
1960-1969
Acta Forestalia Fennica
1953-1968
1933-1952
1913-1932

Articles containing the keyword 'neural net'

Category : Article

article id 5555, category Article
Mauno Pesonen, Arto Kettunen, Petri Räsänen. (1995). Modelling non-industrial private forest landowners’ strategic decision making by using logistic regression and neural networks: Case of predicting the choice of forest taxation basis. Silva Fennica vol. 29 no. 2 article id 5555. https://doi.org/10.14214/sf.a9206
Keywords: logistic regression; Finland; Neural Networks; forest owners; forest taxation; non-industrial private forest landowners,; timber management strategies
Abstract | View details | Full text in PDF | Author Info

In this study, logistic regression and neural networks were used to predict non-industrial private forests (NIPF) landowners’ choice of forest taxation basis. The main frame of reference of the study was the Finnish capital taxation reform of 1993. As a consequence of the reform, landowners were required to choose whether to be taxed according to site-productivity or realized-income during the coming transition period of thirteen years.

The most important factor affecting the landowners’ choice of taxation basis was the harvest rate during the transition period, i.e. the chosen timber management strategy. Furthermore, the estimated personal marginal tax rate and the intention to cut timber during next three years affected the choice. The descriptive landowner variables did not have any marked effect on the choice of forest taxation basis.

On average, logistic regression predicted 71% of the choices correctly; the corresponding figure for neural networks was 63%. In both methods, the choice of site-productivity taxation was predicted more accurately than the choice of realized-income taxation. An increase in the number of model variable did not significantly improve the results of neural networks and logistic regression.

  • Pesonen, E-mail: mp@mm.unknown (email)
  • Kettunen, E-mail: ak@mm.unknown
  • Räsänen, E-mail: pr@mm.unknown

Category : Research article

article id 460, category Research article
Hong Ling, Sandhya Samarasinghe, G. Don Kulasiri. (2009). Modelling variability in full-field displacement profiles and Poisson ratio of wood in compression using stochastic neural networks. Silva Fennica vol. 43 no. 5 article id 460. https://doi.org/10.14214/sf.460
Keywords: wood; digital image correlation; displacement profiles; variability; micro structure; stochastic neural networks; Poisson ratio
Abstract | View details | Full text in PDF | Author Info
Vertical and horizontal displacement profiles in compression parallel-to-grain in a 20 x 20 mm area (30 x 21 or 630 points) in the Tangential–Longitudinal (T–L) and Radial Longitudinal (R–L) sections of small wood columns were obtained from digital image correlation applied to simultaneously captured images of the two surfaces. These consisted of 21 displacement realisations of 30 points along the length of the specimen. They revealed considerable local variations. Stochastic neural networks were successfully developed to simulate trends and noise across and along a specimen in both displacements as well as Poisson ratios in T–L and R–L sections for two selected load levels of 20kN and 40kN. These networks specifically embed noise characteristics extracted from data to generate realistic displacement and Poisson ratio realisations with inherent variability. Models were successfully validated using independent data extracted based on bootstrapping method with high accuracy with R2 ranging from 0.79 to 0.91. The models were further validated successfully using a second approach involving Confidence Intervals generated from the data extracted from the models. Models and experimental results revealed that for 20kN load, both vertical and horizontal displacements in T–L section were less heterogeneous across the specimen (smaller vertical shearing and horizontal strain, respectively) than those in the R–L section. For the 40kN load, both displacement profiles in the T–L section were less noisy and more compact than those for the 20kN load indicating less heterogeneity due to compaction of structure. In the R–L section, larger vertical shearing and horizontal strains persisted at 40 kN load. Poisson ratio decreased with load and it was nonlinear in both sections but T–L section showed much less noise across the specimen than the R–L section.
  • Ling, Centre for Advanced Computational Solutions (C-fACS), Lincoln University, Canterbury, New Zealand E-mail: hl@nn.nz
  • Samarasinghe, Centre for Advanced Computational Solutions (C-fACS), Lincoln University, Canterbury, New Zealand E-mail: sandhya.samarasinghe@lincoln.ac.nz (email)
  • Kulasiri, Centre for Advanced Computational Solutions (C-fACS), Lincoln University, Canterbury, New Zealand E-mail: gdk@nn.nz
article id 309, category Research article
Sandhya Samarasinghe, Don Kulasiri, Tristan Jamieson. (2007). Neural networks for predicting fracture toughness of individual wood samples. Silva Fennica vol. 41 no. 1 article id 309. https://doi.org/10.14214/sf.309
Keywords: Pinus radiata; New Zealand; video imaging; strain energy release rate; Neural Networks; fracture toughness
Abstract | View details | Full text in PDF | Author Info
Strain energy release rate (GIc) of Pinus radiata in the TL opening mode was determined using the compliance crack length relationship. A total of 123 specimens consisting of four sizes of specimen with each size having four different crack lengths were tested. For each specimen, grain and ring angles, density and moisture content were measured. Video imaging, was used to measure crack length during propagation. Since cracks extended in stages, full compliance-crack length relationship was developed for each specimen based on their initial and subsequent crack lengths. No significant differences in GIc, between initial and subsequent crack lengths were found for the smaller specimens by paired sample t-tests, but differences were significant for the largest specimen size. The Average fracture toughness was calculated from GIc and it was 215 kPa.m0.5. Three artificial neural networks were developed to predict the: 1) force required to propagate a crack, 2) crack extension, and 3) fracture toughness of an individual specimen. Each was successful, producing respective R2 of 0.870, 0.865, and 0.621 on validation data. A sensitivity analysis of the networks revealed that the crack length was the most influential with 21% contribution followed by grain angle with 14% contribution for predicting the applied force. This was followed by volume and physical properties. For predicting the crack extension, density had the greatest contribution (20%) followed by previous crack length and force contributing 16% equally. Fracture toughness was dominated by the dimensional parameters of the specimen contributing (42%) followed by anisotropy and physical properties.
  • Samarasinghe, Centre for Advanced Computational Solutions (C-fACS), Lincoln University, New Zealand E-mail: ss@nn.nz (email)
  • Kulasiri, Centre for Advanced Computational Solutions (C-fACS), Lincoln University, New Zealand E-mail: dk@nn.nz
  • Jamieson, Centre for Advanced Computational Solutions (C-fACS), Lincoln University, New Zealand E-mail: tj@nn.nz
article id 618, category Research article
Steen Magnussen, Paul Boudewyn, Mike Wulder, David Seemann. (2000). Predictions of forest inventory cover type proportions using Landsat TM. Silva Fennica vol. 34 no. 4 article id 618. https://doi.org/10.14214/sf.618
Keywords: neural net; maximum likelihood classification; agreement of predictions
Abstract | View details | Full text in PDF | Author Info
The feasibility of generating via Landsat TM data current estimates of cover type proportions for areas lacking this information in the national forest inventory was explored by a case study in New Brunswick. A recent forest management inventory covering 4196 km2 in south-eastern New Brunswick (the test area) and a coregistered Landsat TM scene was used to develop predictive models of 12 cover type proportions in an adjacent 4525 km2 region (the validation area). Four prediction models were considered, one using a maximum likelihood classifier (MLC), and three using the proportions of 30 TM clusters as predictors. The MLC was superior for non-vegetated cover types while a neural net or a prorating of cluster proportions was chosen for predicting vegetated cover types. Most predictions generated for national inventory photo-plots of 2 x 2 km were closer to the most recent inventory results than estimates extrapolated from the test area. Agreement between predictions and current inventory results varied considerably among cover types with model-based predictions outperforming, on average, the simple spatial extensions by about 14%. In this region, an 11-year-old forest inventory for the validation area provided estimates that in half the cases were closer to current inventory estimates than predictions using the optimal Landsat TM model. A strong temporal correlation of photo-plot-level cover type proportions made old-values more consistent than predictions using the optimal Landsat TM model in all but three cases. Prorating of cluster proportions holds promise for large-scale multi-sensor predictions of forest inventory cover types.
  • Magnussen, Canadian Forest Service, 506 West Burnside Road, Victoria B.C., Canada V8Z 1M5 E-mail: smagnussen@pfc.forestry.ca (email)
  • Boudewyn, Canadian Forest Service, 506 West Burnside Road, Victoria B.C., Canada V8Z 1M5 E-mail: pb@nn.ca
  • Wulder, Canadian Forest Service, 506 West Burnside Road, Victoria B.C., Canada V8Z 1M5 E-mail: mw@nn.ca
  • Seemann, Canadian Forest Service, 506 West Burnside Road, Victoria B.C., Canada V8Z 1M5 E-mail: ds@nn.ca

Register
Click this link to register to Silva Fennica.
Log in
If you are a registered user, log in to save your selected articles for later access.
Contents alert
Sign up to receive alerts of new content
Your selected articles