Artikkelit jotka sisältää sanan 'logistic regression'
Kunnostusojitustarpeen ennustaminen ojitusalueilla.
English title: Predicting the need for ditch network maintenance in drained peatland sites in Finland.
Tiivistelmä |
Näytä lisätiedot
|
Artikkeli PDF-muodossa |
Tekijät
Logistic regression models were developed to predict the condition of ditch networks in drained peatland sites in Finland. The data consisted of observations from two forest inventories in which the need for ditch network maintenance had been assessed in the field by classifying the condition of the ditches in the sample stands. In the analysis an indicator variable which referred to one of two condition categories (in need of repair – not in need of repair) was used as the response variable. According to the results, the probability of being in the poor condition category was higher in sites where the time elapsed since drainage was longer, the geographic location was more northern, peat thickness was greater, and plot inclination was smaller. At a probability level of 0.5, the models predicted the category correctly in 69% of the sites in the modeling data, on average. The models were applied to a growth simulator to study the effect of poor drainage conditions on stand-level growth forecasts.
-
Hökkä,
Finnish Forest Research Institute, Rovaniemi Research Station, P.O. Box 16, FIN-96301 Rovaniemi, Finland
Sähköposti:
ei.tietoa@nn.oo
- Alenius, Sähköposti: ei.tietoa@nn.oo
- Salminen, Sähköposti: ei.tietoa@nn.oo
Click this link to register to Suo - Mires and peat.
Jos olet rekisteröitynyt käyttäjä, kirjaudu sisään tallentaaksesi valitsemasi artikkelit myöhempää käyttöä varten.
Valitsemasi artikkelit
Suo - Mires & Peat
