题目

Grade: 34.00 / 40.00

Tips

全让AI生成的

Feedback

=== Task A ===

• Q1 — Zhiping Liang

  The outlier date was not removed, leading to an incorrect result.

• Q2 — Jiameng Wei
• Q3 — Jiameng Wei
• Q4 — Jae

  Incorrect topics needed to have been removed.

• Q5 — Jae

  5A could have included variations of shoot and/or disappointment, as well as their respective synonyms via regular expression patterns.

=== Task B ===

• Part A — Kaixun
• Part B — Kaixun

=== Task C ===

• Q1 — Ahmad

  Need to add 0 for days 25–29 for the user kegan61438051, and handle the missing value for the user dev_discourse to better compare daily trend of each user.

• Q2 — Ahmad
• Q3 — Leila
• Q4 — Leila

  It would be better if you include explanation and interpret the outcome of the categorization after showing the result.

• Q5 — Kai

  The thresholds to detect spammers should use percentiles rather than fixed values for better generalization, and further explain why the fixed values were chosen.

• Q6 — Saher

  • Sentiment scoring is missing. You need to calculate and interpret sentiment values for the dataset.
  • User consistency analysis is missing. There is no evaluation of how individual users’ sentiments or behaviors vary over time.

=== Task D ===

• Q1 — Yunshu

  Well Done!

• Q2 — Jesmin

  Could have clear answer justification.

• Q3 — Conghui

  Discussion of the important variable is needed to provide.

• Q4 — Saher

  Model improvement methods mentioned: 1

• Q5 — Chris

  Ranked in the top 25% (Received full marks)