Notice
Recent Posts
Recent Comments
Link
일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | ||||||
2 | 3 | 4 | 5 | 6 | 7 | 8 |
9 | 10 | 11 | 12 | 13 | 14 | 15 |
16 | 17 | 18 | 19 | 20 | 21 | 22 |
23 | 24 | 25 | 26 | 27 | 28 |
Tags
- distributed training
- Pararel Computing
- root of the word
- NLP
- Data Distribution
- online machine learning
- plotting
- Apache Beam
- Naive Bayes example
- pandas
- level of negative
- ML
- Learn from mistake
- Deep Learning
- variance bias
- negotiate
- Insight
- productivity
- bias vs variance
- NO TYPO
- level of emotion
- sweet spot in fitting
- color theory
- API
- level of positive
- plot color
- machine learning
- utilize
- color and data
- tensorflow
Archives
- Today
- Total
목록color theory (1)
How much am I expressed through the data

Express the level of emotions in either positive or negative through the level of some color feature as following. (i.e) positive : light green -> green -> dark green negative : light red -> red -> dark red #Copyrights:This screenshot is part of the Coursera's Natural Language Processing Specialization by DeepLearning.ai
Data Science/TODOs
2021. 12. 23. 06:12