I am currently working to build an IBM Watson chatbot and I am running into 2 issues.
Overview: I am building a chatbot that asks questions about the user so the bot can then spit out a "decision" or solution based on how the user answers the questions. Ex. Q1:A Q2:B Q3:D Q4:A CHat bot would spit out the responsibilities associated with collecting A,B,D,A. ABDA = Outcome 5 (out of 16)
Problem 1: Two of my questions lead to the user to input a number as their response. (asking for age and another question asks for years of experience) I do not know how to get Watson to know the difference between them without asking/prompting the user to add "years old" or "years Exp". Has anyone ran into issues of Watson not being able to handle numbers through multiple questions.
Problem 2: Anyone have any suggestions on how to get watson to digest all the questions asked then spit out the correct outcome?
Related
I am doing some research on potential options for building a chatbot. I am currently evaluating amazon Lex. The requirements for the bot are quite simple, a user can ask where to find something, the bot will tell them where in a document they will find the answer. All of these questions and answers have already been captured manually so we can easily have an excel sheet with question and answer.
Is there some way to input these pre-defined questions and responses into Lex? From my research I am having a hard time finding any info on something this basic. It won't really require any back and forth between the user and bot, (for ex. User: 'I need to order flowers' Bot: 'What kind of flowers?" etc.)
I have seen some info on incorporating Kendra, but I don't think the requirement is sophisticated enough to warrant using it
Ideally I would love to just hardcode it and say this is a question, and this is the response that should be given. Maybe this use case does not need something as powerful as Lex?
Lex can solve your problem at a fraction of the cost of Kendra.
Having said that, Kendra would be easier to work with when compared to Lex.
If you're got some Python capabilities I would recommend you take a look at the ExcelLexBot repo on GitHub. It is a Serverless Application that reads input from an Excel spreadsheet to build up a basic Lex bot for you.
Closed. This question needs debugging details. It is not currently accepting answers.
Edit the question to include desired behavior, a specific problem or error, and the shortest code necessary to reproduce the problem. This will help others answer the question.
Closed 3 years ago.
Improve this question
My team has published an app that retrieves users' Google Fit data using the REST API to display how active they are and grade their fitness level. I'm interested in the steps and active_minutes metrics.
About 50% of our users have either no data sources or limited data sources. Some of these users are on my team and I have verified they have accepted the required scopes and that the Fit app is indeed recording data.
To test users, I am refreshing their access token (which works fine) and then calling this endpoint to retrieve a list of available data sources:
GET fitness/v1/users/me/dataSources (https://developers.google.com/fit/rest/v1/reference/users/dataSources/list)
Sometimes the datasources will be an empty array, and sometimes it will have a very limited number of data sources (ex. data sources involving calories, but not steps, even though steps are showing up just fine in Fit).
I am requesting the following scopes:
fitness.activity.read
fitness.body.write
The other 50% of our users work just fine and I am completely stumped at what is different about the users that appear to have no (or limited) data sources.
I found that some users had syncing disabled for Google Fit. To enable syncing you must go into the Settings app of your Android phone:
Settings -> Accounts -> Select your Google account being used for Google Fit -> Ensure that Google Fit is enabled for syncing.
This answers the most puzzling case where we had users with some data sources, but not all of the expected data sources such as estimated steps. I presume the limited data sources available were from a short time period before syncing was disabled, or while the user was using an outdated version of Google Fit. I'm also finding users with more obvious problems, like using OAuth with a Google account different than the one connected to Google Fit.
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 11 months ago.
Improve this question
For a while now I have been trying to see if I can get a Whatsapp bot running, in a similar fashion to how Telegram bots work.
I found quite some ambiguous sources or half baked projects that are supposed to be provide Whatsapp bot functionality, but in fact nothing seems to be actually working.
To my knowledge, Whatsapp (unlike Telegram) does not have a public API or openly documented protocol. Does anyone know if it is even possible at all to "automate" communication through Whatsapp, as in send and receive messages, and how? Or would anyone have any good leading points to start developing one myself?
[ https://umstechlabs.com/ai-chatbot/]
The terms WhatsApp Chatbots, WhatsApp Bot, Messenger Bot, and Chatbot have one thing in common — they are all ways to refer to a Bot. Wikipedia provides a great definition for bots: “A Chatterbot, Chatbot, or simply Bot is a text-based dialogue system, which allows you to chat with a technical system.
Bots are everywhere and businesses are changing towards micro apps and cognitive bots for B2B andB2C. There are a few libraries to make your own WhatsApp Bot. A small python framework to create a whatsapp bot, with regex-callback. WhatsApp bot: Send message to a large list of numbers using whatsapp web.
Whatsapp bot without coding. With Xenioo you can create your bot Visually, boost it with AI and Integrations.
try AutoResponder for WA by TK Studio
https://play.google.com/store/apps/details?id=tkstudio.autoresponderforwa&hl=en
it can make conversations using DialogFlow which is support Webhook, and Webhook will send the message to a URL via POST, and from there you can save the message or specific variable to the database and respond it back to DialogFlow.
or you just use the AutoResponder for simple conversations.
Well, when it comes to comparing Whatsapp(chat)bots to Telegram bots there is a huge gap of availability, in the terms of that Telegram is open source and there is an endless count of varied bots of choice made by the community, aside from whatsapp bots mainly developed by official or affiliated companies of whatsapp.
So I haven't found anthing yet more useful bot other than different tastes of chatbots, auto replyers and message schedulers.
At the other side, some Telegram bots are capable of sending specific web searches within the chat, news feeds, reminders, games (and not mentioning the useful bots for groups) etc.
I have been doing some research on this topic lately and there actually are ways you can bridge Telegram messages to Whatsapp.
If you might be interested I found 2 ways of doing this (there may be more); bridging by code with Matterbridge or going the easy way with thirs-party API and automation services. Dropped some links below.
shorturl.at/klFR7
shorturl.at/loBEF
Haven't done it myself so not sure if bot functionality may be affected on chat at the other end though.
For my graduation project I have to do research about defining the efficiency of our chatbot.
The chatbot asks the user to login before starting to chat to the bot. The way that the efficiency is defined at the moment is by deviding the answered questions by the total number of asked questions.
Now our main customer is complaining that the question about the login are also taken in consideration with the efficiency of the chatbot.
I was wondering if someone how to handle this. Is it correct to add the login questions to the efficiency or should is be taken out of consideration or weigh less heavy on the efficiency?
An example of a conversation
I think it should not be considered.
Because it is a key phase where the chat bot start to interact with the user. What is the function of chat bot? -- it is to have a conversation with the users.
Suppose i will call you on the phone and say hello to start a conversation, will you consider hello as a conversation. it is just a phase or trigger to start conversation.
(This is just my opinion )
conclusion, it is just a lamp rubbing action for genie to come out of the lamp. so do not consider the login phase/questions.
I am trying to build a chatbot using Rasa. I have created a basic chatbot by following the steps given in documentation here. I have installed both rasa core and rasa nlu but for now i am using only rasa core as i don't need to extract any information from input.
I have added around 20 intents and their corresponding actions. But when i trying to get response it is recognising 14 intent accurately (tested even by jumbling word, by using synonyms) but for rest 6 intent it is always returning wrong response even if i enter same input as defined in intent.
At first i used spaCy + sklearn pipeline but now i am using sklearn + MITIE but still not getting accurate responses.
Is there any way to improve the accuracy of chatbot.
The best way to improve accuracy is to optimize your bot iteratively based on what conversational analytics tells you about how it interacts with users over time. It's unrealistic to expect a bot to be accurate out of the box, no matter how much NLP you bring to the table.
There are a bunch of tools available for this, with some being stronger in analytics (in addition to health metrics) than others. (I work for one called Chatbase, which is free to use and works with any bot platform.)