CSCI 3387
Topics in Computational intelligence: Machine Learning Projects · 1 instructor on record · 5 sections
Boston College's ratings are bunched at the top — the median instructor averages about 88 out of 100 on the raw scale, and 44% of them clear 90 — so a raw score cannot separate anybody. Every figure here is ranked against BC classes and that rank is re-expressed out of 100, which is what turns a raw 88 — the middle of the university — into a figure that says where somebody actually stands. On this scale 77 is the median, 84 the top quarter and 91 the top tenth. Workload is then an adjustment on top of that, worth nothing at all for an ordinary one. Lighter work than 19% of BC classes.
The class itself, averaged across everyone who taught it — who taught it is the board below.
How is this worked out?
That is what a student picking a section is actually choosing between: the board carries each instructor's own score for this class, and the score above is the class averaged across all of them.
Boston College's own “intellectually challenging” question. It is not a measure of difficulty, grading or workload.
How is this worked out?
Boston College's own “intellectually challenging” question for this class only — students answer it positively about courses that stretched them, so it measures intellectual challenge and NOT grading, workload or quality. Almost every BC class scores high on it, which is why it is shown as a rank rather than a score.
One of Boston College's five workload bands, not a number of hours. Higher means more work, not worse.
How is this worked out?
Boston College reports workload for this class only as one of five bands, not as a number of hours, and it does not publish what the bands mean — so this site will not translate it into hours. Higher means more work, which is not the same as worse.
In this project based class, we will introduce several machine learning concepts, and illustrate and practice their use. These topics will, tentatively, include: classification, data processing, dimensionality reduction, model evaluation and tuning, ensemble learning, regression, clustering, multi layer artificial neural networks and their use for classification, regression, generative adversarial networks, and reinforcement learning.
| Instructor | Curve | Easiness | Challenge |
|---|---|---|---|
| Jose Bento Ayres PereiraComputer Science | 5259 raw · 5 records | 66−3 to the score | 57th |
There is no ranking to read here: Jose Bento Ayres Pereira is the only instructor with evaluation records for CSCI 3387, so the number is a record, not a position.
Also taught by Jose Bento Ayres Pereira
3 other classesOther CSCI classes
All of themThe line
All 13 items · means across 5 recordsBars are ranks, not raw scores. Nothing here is graded good or bad.
What each item asks
- Explains Clearly
- The instructor was effective at creating an understanding of difficult subject matter.
- Available
- The instructor was available for help outside of class.
- Motivating
- The instructor motivated me to do my best work.
- Sparks Interest
- The course stimulated my interest in the subject matter.
- Organized
- The course was well organized.
- Clear Goals
- The learning objectives for the course were clear.
- Useful Later
- What I learned is applicable beyond this course.
- Assignments
- Assignments and exams contributed to my learning.
- Fast Feedback
- I received feedback on my work in a timely manner.
- Useful Feedback
- The feedback I received on my work was meaningful.
- Workload
- How much work the course took outside class, on Boston College's five-band scale — the band boundaries are not published, so this is not a number of hours.
- Challenge
- The course was intellectually challenging.
- Attendance
- Attending class was necessary to do well in this course.
Why these numbers are curved
Both the number and the bar are the rank against every comparable record we hold, printed back out of 100. BC's questions all have a median around 4.5, so on the raw scale every cell here would read in the high eighties and every bar would be nearly full — thirteen numbers that tell you nothing. Each cell prints its raw BC figure and its percentile underneath, so nothing is hidden by the transform. Every bar is one ink colour — nothing here is graded green or red. The hatched bars are not ratings at all: a higher number on either describes a different course, not a better one.
Game log
5 sections · newest first| Instructor | Term | Sec | Instr | Course | Work |
|---|---|---|---|---|---|
| Jose Bento Ayres Pereira | Spring 2026 | 01 | 49 | 53 | 60 |
| Jose Bento Ayres Pereira | Spring 2025 | 01 | 74 | 74 | 80 |
| Jose Bento Ayres Pereira | Fall 2024 | 01 | 67 | 63 | — |
| Jose Bento Ayres Pereira | Spring 2024 | 01 | 51 | 54 | 60 |
| Jose Bento Ayres Pereira | Spring 2023 | 01 | 54 | 54 | 60 |
How to read this board
Instructors are ranked on their record in this class only, never on their average across everything they teach. Scores are out of 100 — Boston College's 1–5 agreement scale, ×20, the way these numbers were printed for years. Workload is BC's own five-band figure, not a number of hours, and is shown as a rank. A dash means the question was not reported, and is never a zero.