Singular Matrix 'a' in solve

Question

I am trying to use predict function in RStudio on some new data. However I keep getting an error, "Error in qr.solve(qr.R(qrX)[p1, p1]) : singular matrix 'a' in solve". Where is my mistake? Should I have a simpler model?

lm.price.new <- lm(price ~ bedrooms * bathrooms * sqft_living * sqft_lot * floors *
                     waterfront * view * condition * grade, data = kc_house_data)
summary(lm.price.new)

new.house <- data.frame(bedrooms = 4, bathrooms = 2, sqft_living = 2560, sqft_lot = 7650, 
                        floors = 1.5, waterfront = 1, view = 3, condition = 5,
                        grade = 10)
predict(lm.price.new, newdata = new.house, interval = "predict")

Probably replace the * with +

Edward
– Edward

2025-02-24 05:30:57 +00:00
Commented Feb 24 at 5:30 — Edward
– Edward, Commented Feb 24 at 5:30

stella1897 · Accepted Answer · 2025-02-24 05:30:24Z

1

I changed the model to not include cross effects as that was what was causing the singularities. Here is what I changed.

lm.price.new1 <- lm(price ~ bedrooms + bathrooms + sqft_living + sqft_lot + floors +
                     waterfront + view + condition + grade, data = kc_house_data)

answered Feb 24 at 5:30

stella1897

576 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Roland Feb 24 at 6:27

I expect that there is strong collinearity between your predictors. You should look into that. You might be better suited with a LASSO model.

Collectives™ on Stack Overflow

Singular Matrix 'a' in solve

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related