GSoC'23 - Week 3 Report

[Baole Fang]

---
title: "Week 3"
date: 2023-06-15
---

### Model training

Basic [model training](
https://github.com/baolef/libreoffice-ci/blob/main/train.py) pipeline is
completed with [testselect](
https://github.com/baolef/libreoffice-ci/blob/main/models/testselect.py)
model. Further optimization is needed to reduce memory and time cost,
together with performance.

Currently, [testselect](
https://github.com/baolef/libreoffice-ci/blob/main/models/testselect.py) is
trained on a subset of size 16384 (containing training and testing set) of
the full dataset of size 122019 due to memory cost, and it has reached a
failure recall of 91.4% and saving 90% of unit test computational cost. Its
detailed confusion matrix is shown below:

|               | Fail (Predicted) | Pass (Predicted) |
|---------------|------------------|------------------|
| Fail (Actual) | 480              | 45               |
| Pass (Actual) | 556910           | 5045893          |
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice/attachments/20230625/4dc0c302/attachment.htm>