> Jaidev is suggesting that fit_intercept=False makes no sense if the data is > sparse. But I think that depends on your target variable. It can make sense **not** to fit intercept e.g. if it has no impact on perf it is faster to optimize without one