Machine Learning Software Resources

Kraken is secure. We store data on secure servers on Amazon's EC2 and S3 services using Amazon's Security Groups and S3 security policies. Access to the servers is controlled using public-key encryption. We use SSL throughout the app, on all pages, to secure every last transaction and piece of data, for all accounts. In terms of security, we use a similar approach to many SaaS solutions that require users to have access to an account in order to see any data. We use a layered approach combining access filters that run every time something is accessed as well as database queries that require users to have access to the account in order for data to be returned. Finally, in terms of our internal processes and policy, we don't have easy access to your projects and data (ie. I can't simply log in as you to see what's in your account), and we only dig into the logs or the database when specifically required for troubleshooting. We don't resell or reuse your data in any way, nor do we have plans to. You can find more information in our Privacy Policy: https://www.bigsquid.com/policy

Show More

Show Less

“Overfitting” means that a model is overly complex and as a result, is unreliable for predicting new data. Overfitting tends to happen when there are too many Drivers relative to the number of data points available. For example: you may only have 50 rows of data and 100 Drivers (columns) in the dataset. The predictive model can use all of the Drivers to come up with a series of complicated rules that perform well against the data used to train the mode, when in reality the predicted Metric may be influenced by only one or two predictors. As a rule of thumb, simpler is better. The more Drivers introduced to a model, the more error exists that can potentially cover up the true underlying relationship that you want to uncover.

Show More

Show Less

It all starts with the data – garbage in, garbage out. Some factors to consider that may have an impact on model quality: - If there is data that’s unclean or unreliable, consider removing it from the set. Unclean or unreliable could mean that a majority of column values are null, high concentration of one value in a single column (i.e. you have a column with the values of 'red', 'green', 'blue' and 90% of the values in the column are 'red'), values in a column are highly unique - If the nature of the data you are gathering has experience a significant change (for example, a major policy change that goes into effect may mean the previous data doesn't resemble new data) - A larger volume of data tends to produce more reliable models, so any additional relevant data points will help, whether those are new observations gathered as time passes, or historical ones gathered from a previously untapped source. If your model still isn’t scoring well, it may also be because the metrics that truly have a relationship with the predicted Metric are not yet captured in the dataset. It could be time to brainstorm what other things might have an effect on the predicted metric, and see if that data can be gathered and included in the dataset. Remember – the predictive algorithms can only identify patterns that are there to be found!

Show More

Show Less