Boost modes: Swift & Express¶
When files do not contain additional case profiling information, n-Infinite offers two modes designed to amplify cases using solid training profiles or demographic data only.
Swift mode¶
Swift mode provides amplification equal to or less than 100% of cases (typically around 50%) with high-quality source profiling, by duplicating that profiling and re-predicting those cases. It is recommended for B2B projects with a scarcity of prediction profiles.
Because machine-learning training is not absolutely exact for each case — some randomness is involved in the selection of training and validation cases — Swift mode introduces minor variations. These variations add diversity while maintaining the distribution of each quota, by capturing the relationships between profiles and responses across the entire training data.
The basic selection is simply the questionnaire's starting question: the system uses all preceding columns for profiling. The remaining parameters relate to quota organization, which can be done at this step or after training and prediction complete.
Boost segmentation¶
Choose the desired quota distribution — up to the maximum available cases — by selecting one or more variables and generating their combinations.
Balance options¶
Keeps the quota distribution harmonious across combined targets. For example, with quotas of 50% men / 50% women and 50% under 30 / 50% over 30, balancing avoids degenerate allocations that technically meet each target but concentrate them:
| Men | Women | |
|---|---|---|
| Under 30 (unbalanced) | 50 | 0 |
| Over 30 (unbalanced) | 0 | 50 |
| Under 30 (balanced) | 25 | 25 |
| Over 30 (balanced) | 25 | 25 |
SmartQuotas¶
SmartQuotas is an exclusive n-Infinite feature for quota scenarios — when the gross volume of predictions exceeds the required quotas. It automatically selects which responses to include in the final file, not only to meet the quotas but also optimizing the similarity of response distributions against the field data for selected questions. Analysts can pre-configure expected distributions based on their professional evaluation of the field data.
You can choose to balance across all questions, or select the key columns where keeping proportions balanced is critical.
Express mode¶
Express mode targets large boosts (1× to n× the prediction rows) and B2C projects. Unlike Swift — which reuses each case's own profiling — Express imputes real profiles from national census data, with pre-loaded country options available. Census column mapping is required, and post-weighting is recommended.
(This section will be expanded with the step-by-step screens.)