@KidneyCancer To all who have tweeted here....I am Dr. Wood's older brother, Ken. Besides being brothers and growing up together, Chris and I were best friends. I am so grateful for the positive impact he had on so many. His passing is and will be especially hard for me.
We're analyzing the Black Friday dataset available on Kaggle using Python's #sklearn package. So far, the #RandomForest model yields the lowest RMSE between predicted and actual values of the target variable (customer purchases). #MarketingStrategy#DataScience
When training a #machinelearning binary classifier (i.e., for a given set of #marketing#data, will a prospect become a #customer?), it is misleading (and wrong!) to rely solely on model accuracy. Metrics like precision, recall and F1 scores need to be ca…https://t.co/oo1fMAWMvL
#Machinelearning algorithms such as Recurrent #NeuralNetworks (RNN) are very effective for sequence tasks because they have "memory". A uni-directional RNN takes info from the past to process later inputs. A bidirectional RNN takes context from both the past and the future.
How to determine the size of your training, development, and test data sets to develop #machinelearning models? For very large sets (> 1M samples), 98 percent training, 1 percent development, and 1 percent test works well. #datascience#marketingstrategy#marketinganalytics#data
How do you come up with a #machinelearning classification model that separates the red dots from the blue dots? BTW, this could be a plot of #marketing or #customer segmentation data and you need to correctly predict to which group the customer belongs. A…https://t.co/bw9YuvHSxq
Next Wednesday. 6pm. Machine Learning convo with panelists from Certiltycs, Unspecified, Dude Solutions, Validic & https://t.co/JTY5gqUWZS. You don't want to miss free food, drinks and conversation.
#nextmattersmost#machinelearning#durham#startups
https://t.co/aV2ejXpKjq
We're really excited to be moderating a #machinelearning panel comprised of location practitioners this coming Wednesday at @smashingboxes! https://t.co/2qAzRQJ5CF