BrainJS/brain.js

Proper way to detect sentiments?

Open

#366 建立於 2019年2月20日

在 GitHub 查看
 (1 留言) (0 反應) (0 負責人)TypeScript (14,095 star) (1,099 fork)batch import
help wantedneeds more info

描述

Hello, I'm trying to create a more complicated version of the Sentiment detector, using this example here: https://scrimba.com/p/pVZJQfg/c3n6qfg

However, I'm getting odd results. It seems to work well when there is very little data. But once I put a lot of data, it returns results like "T T T T T T T T T T T T" or some odd words created by mixing input and output.

What can I do to ensure the integrity of the outputs when I have a lot of data?

貢獻者指南

技術棧
javascripttypescriptnodejs
領域
machine learningai
議題類型
research
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-3 hours
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
needs investigation
前置要求
basic JavaScriptneural network conceptsbrain.js basics
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
40
研究方向
This issue is about troubleshooting a sentiment detector using brain.js. The user reports that with large datasets, the output becomes garbled. To investigate, first review brain.js documentation on training recurrent neural networks and data preprocessing. Check if there are any known issues with large datasets in brain.js (e.g., memory constraints or normalization problems). Look at the example linked in the issue to see if there's a preprocessing step missing. Test with a smaller dataset to compare behavior, and consider whether the network architecture (e.g., layers, neurons) needs adjustment for larger data.