<![CDATA[Fire be a problem that could happen anywhere, in office buildings, housing or in public facilities. The process of the fire did not predictable. The current system which detection of is a fire or not , if there was a fire notice will send a message to the owner of the through smartphone. the system can't detect in which the location of fire, because know the location of fire will ease the evacuation process. Based on these problems , there must be a fire detection system might give warning spot location fire incident that human were inside the building immediately evacuation and fire sprinkler can inject water directly to the fire. The process of the determination of recipient point fire use the room temperature obtained from the results of reading sensors lm35 connected with arduino mega as mikrokontroler to be implemented with the methods naive bayes. Sensor LM35 read the value of room temperature in a recurrent manner , if there is a trigger on a system . then system warning there was a fire on a certain location. Trigger obtained from sensors flame to detect there was a fire or not, when there was a fire. sensor flame will send triger to arduino mega and value lm35 will be processed in to the methods naive bayes.Researchers used a method of bayes naive to determine the classification fire. This method chosen because it is one of a good the classification methods, the categorization of point class fire has been set since the beginning. After the research was done , there are several conclusions .First conclusion a detection system locations fire point when tested in training room air-conditioning with the data testing as many as 36 points obtained value 94% accuracy. Conclusions second, when fire detection system used in the air conditioning of 36 points testing get some 86 % accuracy. Third conclusion, the system takes time 0.48 seconds to determines any decision in which the location of the occurrence of fire incident.]]>
Copyrights © 2018
